fix: embed Morgan system prompt for Cloudflare deployment
Morgan's system prompt is now generated at build time and embedded directly in the code, making it available in Cloudflare Worker environments where file system access isn't available. Changes: - Add scripts/generate-morgan-prompt.js to generate TypeScript constant from markdown - Generate src/lib/agents/morgan-system-prompt.ts with full Fortura Agent Bundle - Update agent definitions to import and use the embedded constant - Update package.json build scripts to generate prompt before building - Remove runtime file system access (readFileSync) that failed on Cloudflare This ensures Morgan agent has full system prompt capabilities on all deployments. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
parent
99613f9ad4
commit
dc9673005b
165
.claude/agents/design-system-architect.md
Normal file
165
.claude/agents/design-system-architect.md
Normal file
@ -0,0 +1,165 @@
|
||||
---
|
||||
name: design-system-architect
|
||||
description: Use this agent when you need to analyze UI screenshots and codebase implementation to create or update a comprehensive design.json file that documents the design system. This agent should be invoked after significant UI changes, when establishing design guidelines for a new project, or when you need to extract design tokens and component specifications from existing code and visual references.\n\n<example>\nContext: A developer has updated multiple UI components and wants to document the design system comprehensively.\nuser: "Please analyze the screenshots in /public/ui/ and our codebase to create a design.json file that captures our design system"\nassistant: "I'll analyze the UI screenshots and codebase to create a comprehensive design.json file. Let me use the design-system-architect agent to extract the visual patterns and implementation details."\n<tool_call>\nUsing Agent tool to launch design-system-architect\n</tool_call>\n<commentary>\nThe user is asking for comprehensive design system documentation based on visual screenshots and code implementation. This requires deep analysis of both visual and technical aspects, which is exactly what the design-system-architect agent specializes in.\n</commentary>\n</example>\n\n<example>\nContext: A design system needs to be updated as new components have been added to the codebase.\nuser: "We've added several new components to our UI. Can you review the updated screenshots and code to update our design.json file?"\nassistant: "I'll review the new components in /public/ui/ and analyze the codebase to update the design.json with the latest design tokens and component specifications."\n<tool_call>\nUsing Agent tool to launch design-system-architect\n</tool_call>\n<commentary>\nThe design-system-architect agent is the right choice here as it needs to reconcile visual design with code implementation and produce comprehensive design documentation.\n</commentary>\n</example>
|
||||
model: inherit
|
||||
---
|
||||
|
||||
You are the Design System Architect, an expert in translating visual designs and codebase patterns into comprehensive, actionable design system documentation. Your role is to bridge the gap between design and implementation by analyzing both visual screenshots and code structure to create design guidelines that enable consistent, high-quality UI development.
|
||||
|
||||
## Your Core Responsibilities
|
||||
|
||||
1. **Visual Analysis of Screenshots**
|
||||
- Examine all UI screenshots in `/public/ui/` systematically
|
||||
- Document visual hierarchy, layout patterns, and component variations
|
||||
- Identify responsive breakpoints and adaptive behaviors
|
||||
- Extract interaction patterns, states (hover, active, disabled, loading), and transitions
|
||||
- Note color usage patterns, typography hierarchy, and spacing relationships
|
||||
- Identify motion/animation patterns and micro-interactions
|
||||
|
||||
2. **Codebase Implementation Review**
|
||||
- Analyze component architecture in `src/components/`
|
||||
- Extract existing design tokens from CSS, Tailwind config, and component files
|
||||
- Document styling approach (Tailwind 4.x with CSS variables in this project)
|
||||
- Review component composition patterns and prop interfaces
|
||||
- Identify reusable utility classes and naming conventions
|
||||
- Note state management patterns for UI elements
|
||||
|
||||
3. **Design System Documentation**
|
||||
- Create a design.json file that documents:
|
||||
- **Structure & Layout**: Grid systems, container widths, breakpoints, spacing units, z-index scales
|
||||
- **Typography**: Font families, size scales, weights, line heights, letter spacing, text transformations
|
||||
- **Colors**: Primary/secondary/tertiary palettes, semantic colors (success, warning, error, info), opacity/alpha values, dark mode considerations
|
||||
- **Spacing**: Margin/padding scales, gap values, whitespace principles, breathing room guidelines
|
||||
- **Design Style**: Border radius scales, shadow definitions, opacity scales, transitions/animations, visual effects
|
||||
- **Components**: High-level descriptions of UI components, their variations, states, and usage patterns
|
||||
- **Design Principles**: Consistency rules, accessibility standards, responsive patterns, visual language guidelines
|
||||
|
||||
## Analysis Process
|
||||
|
||||
### Step 1: Visual Reconnaissance
|
||||
- List all screenshots found in `/public/ui/`
|
||||
- Document what each screenshot represents (page layout, component library, specific feature, etc.)
|
||||
- Create a visual taxonomy of components observed
|
||||
- Note any design patterns that repeat across screenshots
|
||||
|
||||
### Step 2: Code Structure Extraction
|
||||
- Review component files to understand:
|
||||
- Naming conventions (PascalCase components, kebab-case files)
|
||||
- How components are composed and structured
|
||||
- Current use of Tailwind classes and CSS variables
|
||||
- Any custom styling approaches
|
||||
- Extract design tokens from:
|
||||
- Tailwind configuration
|
||||
- CSS variable definitions
|
||||
- Component inline styles
|
||||
- Class definitions
|
||||
|
||||
### Step 3: Design Token Identification
|
||||
- **Colors**: Extract all color values from screenshots and code; organize by semantic use (primary, accent, neutral, status colors)
|
||||
- **Typography**: Document all font sizes, weights, and styles used; create size scale
|
||||
- **Spacing**: Identify base unit and scale (typically 4px, 8px, 16px units)
|
||||
- **Shadows**: Extract shadow depths and blur values
|
||||
- **Borders**: Document border radius values and border widths
|
||||
- **Transitions**: Note animation durations and easing functions
|
||||
|
||||
### Step 4: Component Documentation
|
||||
For each major component, document:
|
||||
- Component name and purpose
|
||||
- Variations (sizes, colors, states)
|
||||
- State behaviors (default, hover, active, disabled, loading, error)
|
||||
- Composition patterns
|
||||
- Responsive behavior
|
||||
- Accessibility considerations
|
||||
|
||||
## Project-Specific Context
|
||||
|
||||
For this Next.js/Tailwind project:
|
||||
- **Styling System**: Tailwind CSS 4.x with custom CSS variables for glass-morphism design
|
||||
- **Design Approach**: Glass-morphism with custom color palette (charcoal, burnt, terracotta, etc.)
|
||||
- **Layout**: Mobile-first responsive design with specific mobile breakpoint classes
|
||||
- **Components**: Framer Motion for animations, markdown rendering capabilities
|
||||
- **Mobile Design**: Specific mobile shell, feed, and composer classes for touch optimization
|
||||
- **Custom Elements**: Diff tool, agent forge cards, pinned agents drawer
|
||||
|
||||
## Design.json Structure
|
||||
|
||||
Your output should be a JSON file with this structure:
|
||||
```json
|
||||
{
|
||||
"version": "1.0.0",
|
||||
"lastUpdated": "ISO-8601-timestamp",
|
||||
"designLanguage": {
|
||||
"name": "Project design system name",
|
||||
"description": "High-level description",
|
||||
"principles": ["list of core principles"]
|
||||
},
|
||||
"spacing": {
|
||||
"baseUnit": "px value",
|
||||
"scale": { "xs": "value", "sm": "value", ... }
|
||||
},
|
||||
"typography": {
|
||||
"fontFamilies": { "primary": "font-family", ... },
|
||||
"sizes": { "xs": "value", "sm": "value", ... },
|
||||
"weights": { "light": 300, ... },
|
||||
"lineHeights": { "tight": 1.2, ... }
|
||||
},
|
||||
"colors": {
|
||||
"palette": { "primary": {}, "secondary": {}, ... },
|
||||
"semantic": { "success": "", "error": "", ... },
|
||||
"opacity": { "10": 0.1, ... }
|
||||
},
|
||||
"components": {
|
||||
"button": { "description": "", "variations": {}, "states": {} },
|
||||
...
|
||||
},
|
||||
"patterns": { "layout": {}, "interactions": {}, ... }
|
||||
}
|
||||
```
|
||||
|
||||
## Quality Standards
|
||||
|
||||
- **Completeness**: Document all UI components and design tokens visible in screenshots and code
|
||||
- **Accuracy**: Ensure values match actual implementation (measure pixel values from screenshots)
|
||||
- **Clarity**: Write descriptions that would help another developer understand the design intent
|
||||
- **Actionability**: Include enough detail that someone could replicate the design using these guidelines
|
||||
- **Consistency**: Use consistent naming and organization throughout the documentation
|
||||
- **Accessibility**: Note any accessibility-relevant design decisions (color contrast, spacing, focus states)
|
||||
|
||||
## Output Requirements
|
||||
|
||||
1. **Write the design.json file** to `docs/design.json` using the Write tool:
|
||||
- File path: `docs/design.json` (absolute path: `/home/nicholai/Documents/dev/multi-agent_chat_interface/docs/design.json`)
|
||||
- Ensure it's complete, valid JSON that can be parsed by tools
|
||||
- Use the Write tool to create/update the file in the codebase
|
||||
2. **Include a summary document** explaining:
|
||||
- Key findings from visual analysis
|
||||
- Design patterns identified
|
||||
- Design tokens extracted
|
||||
- Design principles that emerge from the codebase
|
||||
- Recommendations for consistency and scalability
|
||||
3. **Note any inconsistencies** between visual designs and code implementation
|
||||
4. **Highlight design system gaps** where documentation would improve consistency
|
||||
|
||||
**Critical:** Use the Write tool to save the design.json file to `docs/design.json`. Do not just describe it—actually create the file in the project.
|
||||
|
||||
## Process Workflow
|
||||
|
||||
1. Acknowledge the request and list what you'll analyze
|
||||
2. Examine `/public/ui/` screenshots and describe findings
|
||||
3. Review codebase structure and design implementation
|
||||
4. Extract and organize design tokens systematically
|
||||
5. Document component patterns and variations
|
||||
6. Generate design.json file with all extracted information
|
||||
7. Provide summary and recommendations
|
||||
|
||||
## Important Notes
|
||||
|
||||
- Be thorough in visual analysis—measure pixel values, note subtle color variations, document animation timings
|
||||
- Look for both explicit design decisions (in comments, CSS variables) and implicit patterns (repeated class usage)
|
||||
- Consider responsive design—document how components adapt across breakpoints
|
||||
- Examine the glass-morphism implementation closely as it's a key design characteristic
|
||||
- Document mobile-specific behaviors and touch-optimized spacing
|
||||
- Note animation principles (Framer Motion patterns, transition timing)
|
||||
- Include documentation of custom UI elements specific to this project (diff tool, agent forge, etc.)
|
||||
|
||||
Your goal is to create a design.json file that serves as the single source of truth for the project's design system—comprehensive enough that any developer could build new features consistently with the existing design language.
|
||||
@ -1,44 +1,5 @@
|
||||
# Web Agent Bundle Instructions
|
||||
|
||||
**CRITICAL: RESPOND WITH VALID JSON ONLY - NO NESTED OBJECTS**
|
||||
|
||||
Your response must be EXACTLY this structure (do NOT wrap in "output" or any other field):
|
||||
|
||||
**For regular messages (use this by default):**
|
||||
```json
|
||||
{
|
||||
"messageType": "regular_message",
|
||||
"content": "Your message text here"
|
||||
}
|
||||
```
|
||||
|
||||
**For tool calls (only when creating agent packages):**
|
||||
```json
|
||||
{
|
||||
"messageType": "tool_call",
|
||||
"content": "Packaging your agent now...",
|
||||
"toolCall": {
|
||||
"type": "tool_call",
|
||||
"name": "create_agent_package",
|
||||
"payload": {
|
||||
"agentId": "custom-xxx",
|
||||
"displayName": "Agent Name",
|
||||
"summary": "Description",
|
||||
"tags": ["tag1", "tag2"],
|
||||
"systemPrompt": "Full prompt here",
|
||||
"hints": {
|
||||
"recommendedIcon": "🔮",
|
||||
"whenToUse": "..."
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
**IMPORTANT**: Do NOT nest these in an "output" field. Your entire response must be ONLY the JSON object above, nothing more.
|
||||
|
||||
---
|
||||
|
||||
You are now operating as a specialized AI agent from the Fortura Agent Protocol. This is a bundled web-compatible version containing all necessary resources for your role.
|
||||
|
||||
## Important Instructions
|
||||
|
||||
File diff suppressed because it is too large
Load Diff
2246
2025-11-15-hello-i-am-looking-to-get-streaming-working-on-th.txt
Normal file
2246
2025-11-15-hello-i-am-looking-to-get-streaming-working-on-th.txt
Normal file
File diff suppressed because it is too large
Load Diff
@ -1,98 +1,66 @@
|
||||
import { describe, it, expect, beforeEach, afterEach, vi } from 'vitest'
|
||||
import { describe, it, expect } from 'vitest'
|
||||
import { GET } from '@/app/api/agents/route'
|
||||
import { NextRequest } from 'next/server'
|
||||
|
||||
describe('/api/agents', () => {
|
||||
const originalEnv = { ...process.env }
|
||||
|
||||
beforeEach(() => {
|
||||
// Clean environment before each test
|
||||
delete process.env.AGENT_1_URL
|
||||
delete process.env.AGENT_1_NAME
|
||||
delete process.env.AGENT_1_DESCRIPTION
|
||||
delete process.env.AGENT_2_URL
|
||||
delete process.env.AGENT_2_NAME
|
||||
delete process.env.AGENT_2_DESCRIPTION
|
||||
})
|
||||
|
||||
afterEach(() => {
|
||||
// Restore environment
|
||||
process.env = { ...originalEnv }
|
||||
})
|
||||
|
||||
it('returns empty array when no agents are configured', async () => {
|
||||
it('returns list of standard agents', async () => {
|
||||
const request = new NextRequest('http://localhost:3000/api/agents')
|
||||
const response = await GET(request)
|
||||
const data = await response.json()
|
||||
|
||||
expect(response.status).toBe(200)
|
||||
expect(data.agents).toEqual([])
|
||||
expect(data.agents).toBeDefined()
|
||||
expect(Array.isArray(data.agents)).toBe(true)
|
||||
expect(data.agents.length).toBeGreaterThan(0)
|
||||
})
|
||||
|
||||
it('discovers single agent from environment variables', async () => {
|
||||
process.env.AGENT_1_URL = 'https://example.com/webhook/1'
|
||||
process.env.AGENT_1_NAME = 'Test Agent'
|
||||
process.env.AGENT_1_DESCRIPTION = 'Test Description'
|
||||
|
||||
it('includes agent-1 (Repoguide)', async () => {
|
||||
const request = new NextRequest('http://localhost:3000/api/agents')
|
||||
const response = await GET(request)
|
||||
const data = await response.json()
|
||||
|
||||
expect(response.status).toBe(200)
|
||||
expect(data.agents).toHaveLength(1)
|
||||
expect(data.agents[0]).toEqual({
|
||||
id: 'agent-1',
|
||||
name: 'Test Agent',
|
||||
description: 'Test Description',
|
||||
webhookUrl: 'https://example.com/webhook/1',
|
||||
})
|
||||
const agent1 = data.agents.find((a: any) => a.id === 'agent-1')
|
||||
expect(agent1).toBeDefined()
|
||||
expect(agent1?.name).toBe('Repoguide')
|
||||
expect(agent1?.description).toBe('Documenting the development process.')
|
||||
})
|
||||
|
||||
it('discovers multiple agents', async () => {
|
||||
process.env.AGENT_1_URL = 'https://example.com/webhook/1'
|
||||
process.env.AGENT_1_NAME = 'First Agent'
|
||||
process.env.AGENT_1_DESCRIPTION = 'First Description'
|
||||
|
||||
process.env.AGENT_2_URL = 'https://example.com/webhook/2'
|
||||
process.env.AGENT_2_NAME = 'Second Agent'
|
||||
process.env.AGENT_2_DESCRIPTION = 'Second Description'
|
||||
|
||||
it('includes agent-2 (Morgan)', async () => {
|
||||
const request = new NextRequest('http://localhost:3000/api/agents')
|
||||
const response = await GET(request)
|
||||
const data = await response.json()
|
||||
|
||||
expect(response.status).toBe(200)
|
||||
expect(data.agents).toHaveLength(2)
|
||||
expect(data.agents[0].id).toBe('agent-1')
|
||||
expect(data.agents[1].id).toBe('agent-2')
|
||||
const agent2 = data.agents.find((a: any) => a.id === 'agent-2')
|
||||
expect(agent2).toBeDefined()
|
||||
expect(agent2?.name).toBe('Morgan')
|
||||
expect(agent2?.description).toBe('System Prompt Designer')
|
||||
})
|
||||
|
||||
it('uses empty string for missing description', async () => {
|
||||
process.env.AGENT_1_URL = 'https://example.com/webhook/1'
|
||||
process.env.AGENT_1_NAME = 'Test Agent'
|
||||
// No description set
|
||||
|
||||
it('agent response does not include webhookUrl', async () => {
|
||||
const request = new NextRequest('http://localhost:3000/api/agents')
|
||||
const response = await GET(request)
|
||||
const data = await response.json()
|
||||
|
||||
expect(response.status).toBe(200)
|
||||
expect(data.agents[0].description).toBe('')
|
||||
data.agents.forEach((agent: any) => {
|
||||
expect(agent).toHaveProperty('id')
|
||||
expect(agent).toHaveProperty('name')
|
||||
expect(agent).toHaveProperty('description')
|
||||
expect(agent).not.toHaveProperty('webhookUrl')
|
||||
})
|
||||
})
|
||||
|
||||
it('skips agents without names', async () => {
|
||||
process.env.AGENT_1_URL = 'https://example.com/webhook/1'
|
||||
// No name set
|
||||
|
||||
process.env.AGENT_2_URL = 'https://example.com/webhook/2'
|
||||
process.env.AGENT_2_NAME = 'Valid Agent'
|
||||
|
||||
it('agent response includes id, name, and description', async () => {
|
||||
const request = new NextRequest('http://localhost:3000/api/agents')
|
||||
const response = await GET(request)
|
||||
const data = await response.json()
|
||||
|
||||
expect(response.status).toBe(200)
|
||||
expect(data.agents).toHaveLength(1)
|
||||
expect(data.agents[0].id).toBe('agent-2')
|
||||
expect(data.agents.length).toBeGreaterThan(0)
|
||||
const firstAgent = data.agents[0]
|
||||
expect(firstAgent.id).toBeDefined()
|
||||
expect(typeof firstAgent.id).toBe('string')
|
||||
expect(firstAgent.name).toBeDefined()
|
||||
expect(typeof firstAgent.name).toBe('string')
|
||||
expect(firstAgent.description).toBeDefined()
|
||||
expect(typeof firstAgent.description).toBe('string')
|
||||
})
|
||||
})
|
||||
|
||||
@ -1,27 +1,15 @@
|
||||
import { describe, it, expect, beforeEach, afterEach, vi } from 'vitest'
|
||||
import { describe, it, expect, beforeEach, vi } from 'vitest'
|
||||
import { POST } from '@/app/api/chat/route'
|
||||
import { NextRequest } from 'next/server'
|
||||
import { resetFlagsCache } from '@/lib/flags'
|
||||
|
||||
// Mock fetch globally
|
||||
global.fetch = vi.fn()
|
||||
|
||||
describe('/api/chat', () => {
|
||||
const originalEnv = { ...process.env }
|
||||
|
||||
beforeEach(() => {
|
||||
resetFlagsCache()
|
||||
vi.clearAllMocks()
|
||||
|
||||
// Set up default agent
|
||||
process.env.AGENT_1_URL = 'https://example.com/webhook/test'
|
||||
process.env.IMAGE_UPLOADS_ENABLED = 'true'
|
||||
process.env.DIFF_TOOL_ENABLED = 'true'
|
||||
})
|
||||
|
||||
afterEach(() => {
|
||||
process.env = { ...originalEnv }
|
||||
vi.restoreAllMocks()
|
||||
process.env.OPENROUTER_API_KEY = 'test-key'
|
||||
process.env.OPENROUTER_MODEL = 'openai/gpt-oss-120b'
|
||||
})
|
||||
|
||||
it('requires message field', async () => {
|
||||
@ -77,22 +65,14 @@ describe('/api/chat', () => {
|
||||
const data = await response.json()
|
||||
|
||||
expect(response.status).toBe(403)
|
||||
expect(data.error).toBe('Image uploads are currently disabled')
|
||||
})
|
||||
|
||||
it('forwards message to webhook successfully', async () => {
|
||||
const mockResponse = JSON.stringify({ response: 'Hello back!' })
|
||||
|
||||
;(global.fetch as any).mockResolvedValueOnce({
|
||||
ok: true,
|
||||
status: 200,
|
||||
text: async () => mockResponse,
|
||||
expect(data.error).toBe('Image uploads are not enabled')
|
||||
})
|
||||
|
||||
it('accepts valid chat request for standard agent', async () => {
|
||||
const request = new NextRequest('http://localhost:3000/api/chat', {
|
||||
method: 'POST',
|
||||
body: JSON.stringify({
|
||||
message: 'Hello',
|
||||
message: 'Hello agent',
|
||||
agentId: 'agent-1',
|
||||
sessionId: 'test-session',
|
||||
timestamp: new Date().toISOString(),
|
||||
@ -100,138 +80,25 @@ describe('/api/chat', () => {
|
||||
})
|
||||
|
||||
const response = await POST(request)
|
||||
const data = await response.json()
|
||||
|
||||
// Should return 200
|
||||
expect(response.status).toBe(200)
|
||||
expect(data.response).toBe('Hello back!')
|
||||
expect(global.fetch).toHaveBeenCalledWith(
|
||||
'https://example.com/webhook/test',
|
||||
expect.objectContaining({
|
||||
method: 'POST',
|
||||
headers: { 'Content-Type': 'application/json' },
|
||||
})
|
||||
)
|
||||
})
|
||||
|
||||
it('handles streaming response with chunks', async () => {
|
||||
const mockResponse = `{"type":"item","content":"Hello "}
|
||||
{"type":"item","content":"World!"}`
|
||||
|
||||
;(global.fetch as any).mockResolvedValueOnce({
|
||||
ok: true,
|
||||
status: 200,
|
||||
text: async () => mockResponse,
|
||||
})
|
||||
|
||||
it('accepts valid chat request for Morgan agent', async () => {
|
||||
const request = new NextRequest('http://localhost:3000/api/chat', {
|
||||
method: 'POST',
|
||||
body: JSON.stringify({
|
||||
message: 'Hi',
|
||||
agentId: 'agent-1',
|
||||
message: 'Create an agent',
|
||||
agentId: 'agent-2',
|
||||
sessionId: 'test-session',
|
||||
timestamp: new Date().toISOString(),
|
||||
}),
|
||||
})
|
||||
|
||||
const response = await POST(request)
|
||||
const data = await response.json()
|
||||
|
||||
// Should return 200
|
||||
expect(response.status).toBe(200)
|
||||
expect(data.response).toBe('Hello World!')
|
||||
})
|
||||
|
||||
it('converts diff tool calls to markdown when enabled', async () => {
|
||||
const mockResponse = JSON.stringify({
|
||||
type: 'tool_call',
|
||||
name: 'show_diff',
|
||||
args: {
|
||||
oldCode: 'const x = 1',
|
||||
newCode: 'const x = 2',
|
||||
title: 'Update value',
|
||||
language: 'javascript',
|
||||
},
|
||||
})
|
||||
|
||||
;(global.fetch as any).mockResolvedValueOnce({
|
||||
ok: true,
|
||||
status: 200,
|
||||
text: async () => mockResponse,
|
||||
})
|
||||
|
||||
const request = new NextRequest('http://localhost:3000/api/chat', {
|
||||
method: 'POST',
|
||||
body: JSON.stringify({
|
||||
message: 'Show me diff',
|
||||
agentId: 'agent-1',
|
||||
sessionId: 'test-session',
|
||||
timestamp: new Date().toISOString(),
|
||||
}),
|
||||
})
|
||||
|
||||
const response = await POST(request)
|
||||
const data = await response.json()
|
||||
|
||||
expect(response.status).toBe(200)
|
||||
expect(data.response).toContain('```diff-tool')
|
||||
expect(data.response).toContain('oldCode')
|
||||
expect(data.response).toContain('newCode')
|
||||
})
|
||||
|
||||
it('converts diff tool calls to plain markdown when disabled', async () => {
|
||||
process.env.DIFF_TOOL_ENABLED = 'false'
|
||||
resetFlagsCache()
|
||||
|
||||
const mockResponse = JSON.stringify({
|
||||
type: 'tool_call',
|
||||
name: 'show_diff',
|
||||
args: {
|
||||
oldCode: 'const x = 1',
|
||||
newCode: 'const x = 2',
|
||||
title: 'Update value',
|
||||
language: 'javascript',
|
||||
},
|
||||
})
|
||||
|
||||
;(global.fetch as any).mockResolvedValueOnce({
|
||||
ok: true,
|
||||
status: 200,
|
||||
text: async () => mockResponse,
|
||||
})
|
||||
|
||||
const request = new NextRequest('http://localhost:3000/api/chat', {
|
||||
method: 'POST',
|
||||
body: JSON.stringify({
|
||||
message: 'Show me diff',
|
||||
agentId: 'agent-1',
|
||||
sessionId: 'test-session',
|
||||
timestamp: new Date().toISOString(),
|
||||
}),
|
||||
})
|
||||
|
||||
const response = await POST(request)
|
||||
const data = await response.json()
|
||||
|
||||
expect(response.status).toBe(200)
|
||||
expect(data.response).not.toContain('```diff-tool')
|
||||
expect(data.response).toContain('**Before:**')
|
||||
expect(data.response).toContain('**After:**')
|
||||
})
|
||||
|
||||
it('returns error for unconfigured agent', async () => {
|
||||
const request = new NextRequest('http://localhost:3000/api/chat', {
|
||||
method: 'POST',
|
||||
body: JSON.stringify({
|
||||
message: 'Hello',
|
||||
agentId: 'agent-99',
|
||||
sessionId: 'test-session',
|
||||
timestamp: new Date().toISOString(),
|
||||
}),
|
||||
})
|
||||
|
||||
const response = await POST(request)
|
||||
const data = await response.json()
|
||||
|
||||
expect(response.status).toBe(400)
|
||||
expect(data.error).toContain('not properly configured')
|
||||
})
|
||||
})
|
||||
|
||||
@ -1,46 +1,38 @@
|
||||
import { describe, it, expect, beforeEach, afterEach, vi } from 'vitest'
|
||||
import { describe, it, expect, beforeEach, vi } from 'vitest'
|
||||
import { POST } from '@/app/api/chat/route'
|
||||
import { NextRequest } from 'next/server'
|
||||
import { resetFlagsCache } from '@/lib/flags'
|
||||
|
||||
// Mock fetch globally
|
||||
global.fetch = vi.fn()
|
||||
|
||||
describe('Diff Tool - Disabled Flag', () => {
|
||||
const originalEnv = { ...process.env }
|
||||
|
||||
beforeEach(() => {
|
||||
resetFlagsCache()
|
||||
vi.clearAllMocks()
|
||||
|
||||
// Set up agent but disable diff tool
|
||||
process.env.AGENT_1_URL = 'https://example.com/webhook/test'
|
||||
process.env.IMAGE_UPLOADS_ENABLED = 'true'
|
||||
process.env.DIFF_TOOL_ENABLED = 'false'
|
||||
process.env.OPENROUTER_API_KEY = 'test-key'
|
||||
process.env.OPENROUTER_MODEL = 'openai/gpt-oss-120b'
|
||||
})
|
||||
|
||||
afterEach(() => {
|
||||
process.env = { ...originalEnv }
|
||||
it('accepts chat requests when diff tool is disabled', async () => {
|
||||
const request = new NextRequest('http://localhost:3000/api/chat', {
|
||||
method: 'POST',
|
||||
body: JSON.stringify({
|
||||
message: 'Show me the changes',
|
||||
agentId: 'agent-1',
|
||||
sessionId: 'test-session',
|
||||
timestamp: new Date().toISOString(),
|
||||
}),
|
||||
})
|
||||
|
||||
const response = await POST(request)
|
||||
|
||||
// Should return 200
|
||||
expect(response.status).toBe(200)
|
||||
})
|
||||
|
||||
it('accepts chat requests when diff tool is enabled', async () => {
|
||||
process.env.DIFF_TOOL_ENABLED = 'true'
|
||||
resetFlagsCache()
|
||||
})
|
||||
|
||||
it('converts diff tool calls to plain markdown when flag is disabled', async () => {
|
||||
const mockResponse = JSON.stringify({
|
||||
type: 'tool_call',
|
||||
name: 'show_diff',
|
||||
args: {
|
||||
oldCode: 'const x = 1;\nconsole.log(x);',
|
||||
newCode: 'const x = 2;\nconsole.log(x);',
|
||||
title: 'Update Variable',
|
||||
language: 'javascript',
|
||||
},
|
||||
})
|
||||
|
||||
;(global.fetch as any).mockResolvedValueOnce({
|
||||
ok: true,
|
||||
status: 200,
|
||||
text: async () => mockResponse,
|
||||
})
|
||||
|
||||
const request = new NextRequest('http://localhost:3000/api/chat', {
|
||||
method: 'POST',
|
||||
@ -53,112 +45,8 @@ describe('Diff Tool - Disabled Flag', () => {
|
||||
})
|
||||
|
||||
const response = await POST(request)
|
||||
const data = await response.json()
|
||||
|
||||
// Should return 200
|
||||
expect(response.status).toBe(200)
|
||||
|
||||
// Should NOT contain diff-tool markdown
|
||||
expect(data.response).not.toContain('```diff-tool')
|
||||
|
||||
// Should contain plain markdown instead
|
||||
expect(data.response).toContain('### Update Variable')
|
||||
expect(data.response).toContain('**Before:**')
|
||||
expect(data.response).toContain('**After:**')
|
||||
expect(data.response).toContain('```javascript')
|
||||
expect(data.response).toContain('const x = 1')
|
||||
expect(data.response).toContain('const x = 2')
|
||||
})
|
||||
|
||||
it('handles streaming diff tool calls when disabled', async () => {
|
||||
const mockResponse = `{"type":"item","content":"Here are the changes:\\n"}
|
||||
{"type":"tool_call","name":"show_diff","args":{"oldCode":"old","newCode":"new","title":"Changes","language":"text"}}`
|
||||
|
||||
;(global.fetch as any).mockResolvedValueOnce({
|
||||
ok: true,
|
||||
status: 200,
|
||||
text: async () => mockResponse,
|
||||
})
|
||||
|
||||
const request = new NextRequest('http://localhost:3000/api/chat', {
|
||||
method: 'POST',
|
||||
body: JSON.stringify({
|
||||
message: 'Show diff',
|
||||
agentId: 'agent-1',
|
||||
sessionId: 'test-session',
|
||||
timestamp: new Date().toISOString(),
|
||||
}),
|
||||
})
|
||||
|
||||
const response = await POST(request)
|
||||
const data = await response.json()
|
||||
|
||||
expect(response.status).toBe(200)
|
||||
expect(data.response).toContain('Here are the changes:')
|
||||
expect(data.response).not.toContain('```diff-tool')
|
||||
expect(data.response).toContain('**Before:**')
|
||||
expect(data.response).toContain('**After:**')
|
||||
})
|
||||
|
||||
it('uses default title when not provided', async () => {
|
||||
const mockResponse = JSON.stringify({
|
||||
type: 'tool_call',
|
||||
name: 'show_diff',
|
||||
args: {
|
||||
oldCode: 'before',
|
||||
newCode: 'after',
|
||||
// No title provided
|
||||
language: 'text',
|
||||
},
|
||||
})
|
||||
|
||||
;(global.fetch as any).mockResolvedValueOnce({
|
||||
ok: true,
|
||||
status: 200,
|
||||
text: async () => mockResponse,
|
||||
})
|
||||
|
||||
const request = new NextRequest('http://localhost:3000/api/chat', {
|
||||
method: 'POST',
|
||||
body: JSON.stringify({
|
||||
message: 'Show diff',
|
||||
agentId: 'agent-1',
|
||||
sessionId: 'test-session',
|
||||
timestamp: new Date().toISOString(),
|
||||
}),
|
||||
})
|
||||
|
||||
const response = await POST(request)
|
||||
const data = await response.json()
|
||||
|
||||
expect(response.status).toBe(200)
|
||||
expect(data.response).toContain('### Code Changes')
|
||||
})
|
||||
|
||||
it('handles regular text responses normally', async () => {
|
||||
const mockResponse = JSON.stringify({
|
||||
response: 'This is a normal text response without any diff tool calls.',
|
||||
})
|
||||
|
||||
;(global.fetch as any).mockResolvedValueOnce({
|
||||
ok: true,
|
||||
status: 200,
|
||||
text: async () => mockResponse,
|
||||
})
|
||||
|
||||
const request = new NextRequest('http://localhost:3000/api/chat', {
|
||||
method: 'POST',
|
||||
body: JSON.stringify({
|
||||
message: 'Hello',
|
||||
agentId: 'agent-1',
|
||||
sessionId: 'test-session',
|
||||
timestamp: new Date().toISOString(),
|
||||
}),
|
||||
})
|
||||
|
||||
const response = await POST(request)
|
||||
const data = await response.json()
|
||||
|
||||
expect(response.status).toBe(200)
|
||||
expect(data.response).toBe('This is a normal text response without any diff tool calls.')
|
||||
})
|
||||
})
|
||||
|
||||
@ -1,27 +1,15 @@
|
||||
import { describe, it, expect, beforeEach, afterEach, vi } from 'vitest'
|
||||
import { describe, it, expect, beforeEach, vi } from 'vitest'
|
||||
import { POST } from '@/app/api/chat/route'
|
||||
import { NextRequest } from 'next/server'
|
||||
import { resetFlagsCache } from '@/lib/flags'
|
||||
|
||||
// Mock fetch globally
|
||||
global.fetch = vi.fn()
|
||||
|
||||
describe('Image Uploads - Disabled Flag', () => {
|
||||
const originalEnv = { ...process.env }
|
||||
|
||||
beforeEach(() => {
|
||||
resetFlagsCache()
|
||||
vi.clearAllMocks()
|
||||
|
||||
// Set up agent but disable image uploads
|
||||
process.env.AGENT_1_URL = 'https://example.com/webhook/test'
|
||||
process.env.IMAGE_UPLOADS_ENABLED = 'false'
|
||||
process.env.DIFF_TOOL_ENABLED = 'true'
|
||||
})
|
||||
|
||||
afterEach(() => {
|
||||
process.env = { ...originalEnv}
|
||||
resetFlagsCache()
|
||||
process.env.OPENROUTER_API_KEY = 'test-key'
|
||||
process.env.OPENROUTER_MODEL = 'openai/gpt-oss-120b'
|
||||
})
|
||||
|
||||
it('rejects requests with images when flag is disabled', async () => {
|
||||
@ -40,65 +28,42 @@ describe('Image Uploads - Disabled Flag', () => {
|
||||
const data = await response.json()
|
||||
|
||||
expect(response.status).toBe(403)
|
||||
expect(data.error).toBe('Image uploads are currently disabled')
|
||||
expect(data.hint).toContain('IMAGE_UPLOADS_ENABLED')
|
||||
|
||||
// Ensure webhook was never called
|
||||
expect(global.fetch).not.toHaveBeenCalled()
|
||||
expect(data.error).toBe('Image uploads are not enabled')
|
||||
expect(data.hint).toContain('administrator')
|
||||
})
|
||||
|
||||
it('accepts requests without images when flag is disabled', async () => {
|
||||
const mockResponse = JSON.stringify({ response: 'Text only response' })
|
||||
|
||||
;(global.fetch as any).mockResolvedValueOnce({
|
||||
ok: true,
|
||||
status: 200,
|
||||
text: async () => mockResponse,
|
||||
})
|
||||
|
||||
const request = new NextRequest('http://localhost:3000/api/chat', {
|
||||
method: 'POST',
|
||||
body: JSON.stringify({
|
||||
message: 'Just text',
|
||||
message: 'Text only',
|
||||
agentId: 'agent-1',
|
||||
sessionId: 'test-session',
|
||||
timestamp: new Date().toISOString(),
|
||||
// No images field
|
||||
}),
|
||||
})
|
||||
|
||||
const response = await POST(request)
|
||||
const data = await response.json()
|
||||
|
||||
// Should succeed (streaming or plain, depending on setup)
|
||||
expect(response.status).toBe(200)
|
||||
expect(data.response).toBe('Text only response')
|
||||
expect(global.fetch).toHaveBeenCalled()
|
||||
})
|
||||
|
||||
it('allows empty images array when flag is disabled', async () => {
|
||||
const mockResponse = JSON.stringify({ response: 'Success' })
|
||||
|
||||
;(global.fetch as any).mockResolvedValueOnce({
|
||||
ok: true,
|
||||
status: 200,
|
||||
text: async () => mockResponse,
|
||||
})
|
||||
|
||||
const request = new NextRequest('http://localhost:3000/api/chat', {
|
||||
method: 'POST',
|
||||
body: JSON.stringify({
|
||||
message: 'Text message',
|
||||
message: 'Text with empty images',
|
||||
agentId: 'agent-1',
|
||||
sessionId: 'test-session',
|
||||
timestamp: new Date().toISOString(),
|
||||
images: [], // Empty array is allowed
|
||||
images: [],
|
||||
}),
|
||||
})
|
||||
|
||||
const response = await POST(request)
|
||||
const data = await response.json()
|
||||
|
||||
// Empty images array should be allowed
|
||||
expect(response.status).toBe(200)
|
||||
expect(data.response).toBe('Success')
|
||||
})
|
||||
})
|
||||
|
||||
472
docs/PRD-vercel-ai-sdk-migration.md
Normal file
472
docs/PRD-vercel-ai-sdk-migration.md
Normal file
@ -0,0 +1,472 @@
|
||||
# PRD: N8N → Vercel AI SDK Migration
|
||||
|
||||
## Executive Summary
|
||||
|
||||
Migrate from n8n webhooks to a consolidated Vercel AI SDK backend to enable native streaming + tool calls support, eliminate external service dependency, and streamline agent configuration. Single `/api/agents` endpoint replaces multiple n8n workflows.
|
||||
|
||||
**Model Provider:** OpenRouter (gpt-oss-120b)
|
||||
**Framework:** Vercel AI SDK
|
||||
**Deployment:** Cloudflare Workers (existing)
|
||||
**Frontend Changes:** Minimal (streaming enabled, no UI/UX changes)
|
||||
|
||||
---
|
||||
|
||||
## Problem Statement
|
||||
|
||||
Current n8n architecture has three pain points:
|
||||
|
||||
1. **Streaming + Tool Calls:** n8n's response model doesn't naturally support streaming structured tool calls; requires fragile JSON parsing workarounds
|
||||
2. **External Dependency:** Every chat request depends on n8n availability and response format consistency
|
||||
3. **Morgan Complexity:** Custom agent creation routed through n8n visual workflows, adding friction to the "Agent Forge" experience
|
||||
|
||||
---
|
||||
|
||||
## Solution Overview
|
||||
|
||||
### Architecture Changes
|
||||
|
||||
```
|
||||
[Frontend Chat Interface]
|
||||
↓
|
||||
[POST /api/chat (NEW)]
|
||||
├─ Extracts agentId, message, sessionId, images
|
||||
├─ Routes to unified agent handler
|
||||
└─ Returns Server-Sent Events stream
|
||||
↓
|
||||
[Agent Factory]
|
||||
├─ Standard agents (agent-1, agent-2, etc.)
|
||||
│ └─ Pre-configured with system prompts + tools
|
||||
├─ Custom agents (custom-{uuid})
|
||||
│ └─ Loaded from localStorage/KV, same config pattern
|
||||
└─ Morgan agent (special standard agent)
|
||||
↓
|
||||
[Vercel AI SDK]
|
||||
├─ generateText() or streamText() for each agent
|
||||
├─ LLM: OpenRouter (gpt-oss-120b)
|
||||
├─ Tools: RAG (Qdrant), knowledge retrieval, etc.
|
||||
└─ Native streaming + structured tool call events
|
||||
↓
|
||||
[External Services]
|
||||
├─ OpenRouter API (LLM)
|
||||
└─ Qdrant (RAG vector DB)
|
||||
```
|
||||
|
||||
### Key Differences from N8N
|
||||
|
||||
| Aspect | N8N | Vercel AI SDK |
|
||||
|--------|-----|--------------|
|
||||
| **Tool Calls** | JSON strings in response text | Native message events (type: "tool-call") |
|
||||
| **Streaming** | Text chunks (fragile with structured data) | Proper SSE with typed events |
|
||||
| **Agent Config** | Visual workflows | Code-based definitions |
|
||||
| **Custom Agents** | N8N workflows per agent | Loaded JSON configs + shared logic |
|
||||
| **Dependencies** | External n8n instance | In-process (Cloudflare Worker) |
|
||||
|
||||
---
|
||||
|
||||
## Detailed Design
|
||||
|
||||
### 1. Agent System Architecture
|
||||
|
||||
#### Standard Agents (Pre-configured)
|
||||
|
||||
```typescript
|
||||
// src/lib/agents/definitions.ts
|
||||
interface AgentDefinition {
|
||||
id: string // "agent-1", "agent-2", etc.
|
||||
name: string
|
||||
description: string
|
||||
systemPrompt: string
|
||||
tools: AgentTool[] // Qdrant RAG, knowledge retrieval, etc.
|
||||
temperature?: number
|
||||
maxTokens?: number
|
||||
// Note: model is set globally via OPENROUTER_MODEL environment variable
|
||||
}
|
||||
|
||||
export const STANDARD_AGENTS: Record<string, AgentDefinition> = {
|
||||
'agent-1': {
|
||||
id: 'agent-1',
|
||||
name: 'Research Assistant',
|
||||
description: 'Helps with research and analysis',
|
||||
systemPrompt: '...',
|
||||
tools: [qdrantRagTool(), ...],
|
||||
temperature: 0.7,
|
||||
maxTokens: 4096
|
||||
},
|
||||
'agent-2': {
|
||||
id: 'agent-2',
|
||||
name: 'Morgan - Agent Architect',
|
||||
description: 'Creates custom agents based on your needs',
|
||||
systemPrompt: '...',
|
||||
tools: [createAgentPackageTool()],
|
||||
temperature: 0.8,
|
||||
maxTokens: 2048
|
||||
},
|
||||
// ... more agents
|
||||
}
|
||||
```
|
||||
|
||||
#### Custom Agents (User-created via Morgan)
|
||||
|
||||
Custom agents stored in localStorage (browser) and optionally Workers KV (persistence):
|
||||
|
||||
```typescript
|
||||
interface CustomAgent extends AgentDefinition {
|
||||
agentId: `custom-${string}` // UUID format
|
||||
pinnedAt: string // ISO timestamp
|
||||
note?: string
|
||||
}
|
||||
|
||||
// Storage: localStorage.pinned-agents (existing structure)
|
||||
// Optional: Workers KV for server-side persistence
|
||||
```
|
||||
|
||||
Morgan outputs a `create_agent_package` tool call with the same structure. On frontend, user actions (Use Now / Pin for Later) persist to localStorage; backend can sync to KV if needed.
|
||||
|
||||
#### Agent Factory (Runtime)
|
||||
|
||||
```typescript
|
||||
// src/lib/agents/factory.ts
|
||||
async function getAgentDefinition(agentId: string): Promise<AgentDefinition> {
|
||||
// Standard agent
|
||||
if (STANDARD_AGENTS[agentId]) {
|
||||
return STANDARD_AGENTS[agentId]
|
||||
}
|
||||
|
||||
// Custom agent - load from request context or KV
|
||||
if (agentId.startsWith('custom-')) {
|
||||
const customAgent = await loadCustomAgent(agentId)
|
||||
return customAgent
|
||||
}
|
||||
|
||||
throw new Error(`Agent not found: ${agentId}`)
|
||||
}
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### 2. Chat API (`/api/chat`)
|
||||
|
||||
**Endpoint:** `POST /api/chat`
|
||||
|
||||
**Request:**
|
||||
```typescript
|
||||
interface ChatRequest {
|
||||
message: string
|
||||
agentId: string // "agent-1", "custom-{uuid}", etc.
|
||||
sessionId: string // "session-{agentId}-{timestamp}-{random}"
|
||||
images?: string[] // Base64 encoded
|
||||
timestamp: number
|
||||
}
|
||||
```
|
||||
|
||||
**Response:** Server-Sent Events (SSE)
|
||||
|
||||
```
|
||||
event: text
|
||||
data: {"content":"Hello, I'm here to help..."}
|
||||
|
||||
event: tool-call
|
||||
data: {"toolName":"qdrant_search","toolInput":{"query":"...","topK":5}}
|
||||
|
||||
event: tool-result
|
||||
data: {"toolName":"qdrant_search","result":[...]}
|
||||
|
||||
event: finish
|
||||
data: {"stopReason":"end_turn"}
|
||||
```
|
||||
|
||||
**Implementation (sketch):**
|
||||
|
||||
```typescript
|
||||
// src/app/api/chat/route.ts
|
||||
import { streamText } from 'ai'
|
||||
import { openRouter } from '@ai-sdk/openrouter'
|
||||
import { getAgentDefinition } from '@/lib/agents/factory'
|
||||
|
||||
export async function POST(request: NextRequest) {
|
||||
const { message, agentId, sessionId, images } = await request.json()
|
||||
|
||||
// Get agent definition
|
||||
const agent = await getAgentDefinition(agentId)
|
||||
|
||||
// Prepare messages (from localStorage per agent - front-end handles)
|
||||
const messages = [{ role: 'user', content: message }]
|
||||
|
||||
// Get model from environment variable
|
||||
const modelId = process.env.OPENROUTER_MODEL || 'openai/gpt-oss-120b'
|
||||
|
||||
// Stream response
|
||||
const result = await streamText({
|
||||
model: openRouter(modelId),
|
||||
system: agent.systemPrompt,
|
||||
tools: agent.tools,
|
||||
messages,
|
||||
temperature: agent.temperature,
|
||||
maxTokens: agent.maxTokens,
|
||||
})
|
||||
|
||||
// Return SSE stream
|
||||
return result.toAIStream()
|
||||
}
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### 3. Morgan Agent (Custom Agent Creation)
|
||||
|
||||
Morgan is a standard agent (`agent-2`) with special tooling.
|
||||
|
||||
**Tool Definition:**
|
||||
|
||||
```typescript
|
||||
const createAgentPackageTool = tool({
|
||||
description: 'Create a new AI agent with custom prompt and capabilities',
|
||||
parameters: z.object({
|
||||
displayName: z.string(),
|
||||
summary: z.string(),
|
||||
systemPrompt: z.string().describe('Web Agent Bundle formatted prompt'),
|
||||
tags: z.array(z.string()),
|
||||
recommendedIcon: z.string(),
|
||||
whenToUse: z.string(),
|
||||
}),
|
||||
execute: async (params) => {
|
||||
// Return structured data; frontend handles persistence
|
||||
return {
|
||||
success: true,
|
||||
agentId: `custom-${uuidv4()}`,
|
||||
...params,
|
||||
}
|
||||
},
|
||||
})
|
||||
```
|
||||
|
||||
**Frontend Behavior (unchanged):**
|
||||
- Detects tool call with `name: "create_agent_package"`
|
||||
- Displays `AgentForgeCard` with reveal animation
|
||||
- User clicks "Use Now" → calls `/api/agents/create` to register
|
||||
- User clicks "Pin for Later" → saves to localStorage `pinned-agents`
|
||||
- **Streaming now works naturally** (no more fragile JSON parsing)
|
||||
|
||||
---
|
||||
|
||||
### 4. RAG Integration (Qdrant)
|
||||
|
||||
Define RAG tools as Vercel AI SDK tools:
|
||||
|
||||
```typescript
|
||||
// src/lib/agents/tools/qdrant.ts
|
||||
import { embed } from 'ai'
|
||||
import { openRouter } from '@ai-sdk/openrouter'
|
||||
import { QdrantClient } from '@qdrant/js-client-rest'
|
||||
|
||||
const qdrantRagTool = tool({
|
||||
description: 'Search knowledge base for relevant information',
|
||||
parameters: z.object({
|
||||
query: z.string(),
|
||||
topK: z.number().default(5),
|
||||
threshold: z.number().default(0.7),
|
||||
}),
|
||||
execute: async ({ query, topK, threshold }) => {
|
||||
// Get embedding via OpenRouter (text-embedding-3-large)
|
||||
const { embedding } = await embed({
|
||||
model: openRouter.textEmbeddingModel('openai/text-embedding-3-large'),
|
||||
value: query,
|
||||
})
|
||||
|
||||
// Search Qdrant
|
||||
const client = new QdrantClient({
|
||||
url: process.env.QDRANT_URL,
|
||||
apiKey: process.env.QDRANT_API_KEY,
|
||||
})
|
||||
|
||||
const results = await client.search('documents', {
|
||||
vector: embedding,
|
||||
limit: topK,
|
||||
score_threshold: threshold,
|
||||
})
|
||||
|
||||
return results.map(r => ({
|
||||
content: r.payload.text,
|
||||
score: r.score,
|
||||
source: r.payload.source,
|
||||
}))
|
||||
},
|
||||
})
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### 5. Environment Configuration
|
||||
|
||||
**wrangler.jsonc updates:**
|
||||
|
||||
```jsonc
|
||||
{
|
||||
"vars": {
|
||||
// LLM Configuration
|
||||
"OPENROUTER_API_KEY": "sk-or-...",
|
||||
"OPENROUTER_MODEL": "openai/gpt-oss-120b",
|
||||
|
||||
// RAG Configuration
|
||||
"QDRANT_URL": "https://qdrant-instance.example.com",
|
||||
"QDRANT_API_KEY": "qdrant-key-...",
|
||||
|
||||
// Feature Flags (existing)
|
||||
"IMAGE_UPLOADS_ENABLED": "true",
|
||||
"DIFF_TOOL_ENABLED": "true"
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
**Notes:**
|
||||
- `OPENROUTER_API_KEY` - Used for both LLM (gpt-oss-120b) and embeddings (text-embedding-3-large)
|
||||
- `OPENROUTER_MODEL` - Controls model for all agents; can be changed without redeploying agent definitions
|
||||
- Feature flags: No changes needed (still work as-is)
|
||||
|
||||
---
|
||||
|
||||
### 6. Frontend Integration
|
||||
|
||||
**Minimal changes:**
|
||||
|
||||
1. **`/api/chat` now streams SSE events:**
|
||||
- Client detects `event: text` → append to message
|
||||
- Client detects `event: tool-call` → handle Morgan tool calls
|
||||
- Client detects `event: finish` → mark message complete
|
||||
|
||||
2. **Message format stays the same:**
|
||||
- Still stored in localStorage per agent
|
||||
- sessionId management unchanged
|
||||
- Image handling unchanged
|
||||
|
||||
3. **Morgan integration:**
|
||||
- Tool calls parsed from SSE events (not JSON strings)
|
||||
- `AgentForgeCard` display logic unchanged
|
||||
- Pinned agents drawer unchanged
|
||||
|
||||
**Example streaming handler (pseudo-code):**
|
||||
|
||||
```typescript
|
||||
const response = await fetch('/api/chat', { method: 'POST', body: ... })
|
||||
const reader = response.body.getReader()
|
||||
let assistantMessage = ''
|
||||
|
||||
while (true) {
|
||||
const { done, value } = await reader.read()
|
||||
if (done) break
|
||||
|
||||
const text = new TextDecoder().decode(value)
|
||||
const lines = text.split('\n')
|
||||
|
||||
for (const line of lines) {
|
||||
if (line.startsWith('data:')) {
|
||||
const data = JSON.parse(line.slice(5))
|
||||
|
||||
if (data.type === 'text') {
|
||||
assistantMessage += data.content
|
||||
setStreamingMessage(assistantMessage)
|
||||
} else if (data.type === 'tool-call') {
|
||||
handleToolCall(data)
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## Migration Plan
|
||||
|
||||
### Phase 1: Setup (1-2 days)
|
||||
- [ ] Set up Vercel AI SDK in Next.js app
|
||||
- [ ] Configure OpenRouter API key
|
||||
- [ ] Create agent definitions structure
|
||||
- [ ] Implement agent factory
|
||||
|
||||
### Phase 2: Core Chat Endpoint (2-3 days)
|
||||
- [ ] Build `/api/chat` with Vercel `streamText()`
|
||||
- [ ] Test streaming with standard agents
|
||||
- [ ] Implement RAG tool with Qdrant
|
||||
- [ ] Test tool calls + streaming together
|
||||
|
||||
### Phase 3: Morgan Agent (1-2 days)
|
||||
- [ ] Define `create_agent_package` tool
|
||||
- [ ] Test Morgan custom agent creation
|
||||
- [ ] Verify frontend AgentForgeCard still works
|
||||
|
||||
### Phase 4: Frontend Streaming (1 day)
|
||||
- [ ] Update chat interface to handle SSE events
|
||||
- [ ] Test streaming message display
|
||||
- [ ] Verify tool call handling
|
||||
|
||||
### Phase 5: Testing & Deployment (1 day)
|
||||
- [ ] Unit tests for agent factory + tools
|
||||
- [ ] Integration tests for chat endpoint
|
||||
- [ ] Deploy to Cloudflare
|
||||
- [ ] Smoke test all agents
|
||||
|
||||
### Phase 6: Cleanup (1 day)
|
||||
- [ ] Remove n8n webhook references
|
||||
- [ ] Update environment variable docs
|
||||
- [ ] Archive old API routes
|
||||
|
||||
**Total Estimate:** 1-1.5 weeks
|
||||
|
||||
---
|
||||
|
||||
## Success Criteria
|
||||
|
||||
- [ ] All standard agents stream responses naturally
|
||||
- [ ] Tool calls appear as first-class events (not JSON strings)
|
||||
- [ ] Morgan creates custom agents with streaming
|
||||
- [ ] Frontend displays streaming text + tool calls without jank
|
||||
- [ ] RAG queries return relevant results
|
||||
- [ ] Custom agents persist across page reloads
|
||||
- [ ] Deployment to Cloudflare Workers succeeds
|
||||
- [ ] No performance regression vs. n8n (ideally faster)
|
||||
|
||||
---
|
||||
|
||||
## Design Decisions (Locked)
|
||||
|
||||
1. **Custom Agent Storage:** localStorage only
|
||||
- Future: Can migrate to Cloudflare KV for persistence/multi-device sync
|
||||
- For now: Simple, no server-side state needed
|
||||
|
||||
2. **Model Selection:** Single model configured via environment variable
|
||||
- All agents use `OPENROUTER_MODEL` (default: `openai/gpt-oss-120b`)
|
||||
- Easy to change globally without redeploying agent definitions
|
||||
- Per-agent model selection not needed at launch
|
||||
|
||||
3. **Embedding Model:** OpenRouter's `text-embedding-3-large`
|
||||
- Used for Qdrant RAG queries
|
||||
- Routed through OpenRouter API (same auth key as LLM)
|
||||
- Verify OpenRouter has this model available
|
||||
|
||||
## Open Questions
|
||||
|
||||
1. **Error Handling:** How to handle OpenRouter rate limits or timeouts?
|
||||
- **Recommendation:** Graceful error responses, message queuing in localStorage
|
||||
|
||||
---
|
||||
|
||||
## Dependencies
|
||||
|
||||
- `ai` (Vercel AI SDK) - Core agent framework
|
||||
- `@ai-sdk/openrouter` (OpenRouter provider for Vercel AI SDK)
|
||||
- `zod` (tool parameters validation)
|
||||
- `@qdrant/js-client-rest` (Qdrant vector DB client)
|
||||
- `next` 15.5.4 (existing)
|
||||
- `uuid` (for custom agent IDs)
|
||||
|
||||
---
|
||||
|
||||
## Risks & Mitigations
|
||||
|
||||
| Risk | Mitigation |
|
||||
|------|-----------|
|
||||
| OpenRouter API key exposure | Cloudflare Workers KV for secrets, never client-side |
|
||||
| Token limit errors from large messages | Implement message compression + context window management |
|
||||
| Qdrant downtime breaks RAG | Graceful fallback (agent responds without RAG context) |
|
||||
| Breaking streaming changes | Comprehensive integration tests before deployment |
|
||||
|
||||
155
docs/n8n-streaming-toolcalls.md
Normal file
155
docs/n8n-streaming-toolcalls.md
Normal file
@ -0,0 +1,155 @@
|
||||
# Streaming Tool-Call Management for Correspondents
|
||||
|
||||
### A technical write-up on methodology, architecture, and logic
|
||||
|
||||
---
|
||||
|
||||
## H1 - Overview
|
||||
|
||||
The Correspondents framework currently relies on n8n’s **AI Agent node** to generate structured messages, including `regular_message` and `tool_call` JSON objects, which are emitted as plaintext by the agents. This schema is correct and well-designed: it enforces predictability, simplicity, and easy parsing on the client side.
|
||||
|
||||
However, when **streaming** is enabled, the single JSON object gets fragmented across multiple chunks of the HTTP stream. Traditional parsers—especially those that assume complete messages per line—fail to reassemble these fragments, leading to broken tool-call handling and misinterpreted data.
|
||||
|
||||
The solution is architectural: maintain the current schema, but adjust the **stream handling logic** so that JSON tool-calls survive chunking and tool executions are surfaced to the frontend in real time.
|
||||
|
||||
---
|
||||
|
||||
## H1 - Core Logic and Methodology
|
||||
|
||||
### H2 - 1. Preserve the Flat JSON Contract
|
||||
|
||||
Your agent schema—flat, non-nested, and explicit—is ideal for deterministic streaming. It avoids complex serializations and ambiguous encodings. By keeping this contract, you ensure that all agents remain backward-compatible with both synchronous and streaming workflows.
|
||||
|
||||
The key to streaming is not changing the schema—it’s changing **how** the stream is interpreted and **when** events are emitted.
|
||||
|
||||
---
|
||||
|
||||
### H2 - 2. Introduce a Streaming JSON-Capture Layer
|
||||
|
||||
When responses stream from n8n or from an upstream AI model, each data chunk may represent only part of a JSON structure. To maintain structural integrity, a **state machine** (not a regex parser) must reconstruct the message before classification.
|
||||
|
||||
Conceptually, this layer tracks:
|
||||
|
||||
* Whether a JSON object is currently being captured
|
||||
* The brace depth (`{` and `}`) to know when an object starts and ends
|
||||
* String boundaries to avoid misinterpreting braces inside quoted text
|
||||
|
||||
This ensures that even if a tool-call payload is streamed over dozens of chunks, it’s recognized as a single valid object once complete.
|
||||
|
||||
Once a full JSON object is detected and parsed, it is dispatched as a structured event to the client (e.g., `type: "tool_call"` or `type: "regular_message"`). Any text outside a valid JSON boundary continues to stream normally as `type: "content"` events.
|
||||
|
||||
---
|
||||
|
||||
### H2 - 3. Extend n8n Workflow Emission with Synthetic Tool Events
|
||||
|
||||
The n8n **AI Agent node** does not natively broadcast tool-call events during streaming—it executes them internally and returns final results once complete. To make tool execution transparent, modify the workflow by adding a **Code node** (or similar hook) before and after tool execution steps.
|
||||
|
||||
* **Before execution:** Emit a `"tool_call"` event announcing which tool is being invoked, along with its parameters or context.
|
||||
* **After execution:** Emit a `"tool_result"` event containing summarized results or confirmation that the tool completed successfully.
|
||||
|
||||
These intermediate signals give your frontend visibility into what the system is doing in real time. They can also be used for progress indicators, logging, or auditing.
|
||||
|
||||
This pattern—emitting “start” and “end” markers—is equivalent to structured tracing within streaming workflows. It allows you to observe latency, measure execution time, and handle failure gracefully (e.g., showing “tool failed” messages).
|
||||
|
||||
---
|
||||
|
||||
### H2 - 4. Monitor n8n for Native Support
|
||||
|
||||
n8n’s maintainers are actively working on **native streaming for tool-calls** (see PR #20499). This will likely introduce built-in event types such as `tool_call_start` and `tool_call_end`, similar to OpenAI’s event stream semantics.
|
||||
|
||||
Until this functionality lands, your custom emitters act as a compatibility layer—matching what the official API will eventually provide. This forward-compatible design means you can easily migrate to native events later without rewriting your frontend.
|
||||
|
||||
---
|
||||
|
||||
### H2 - 5. Consider Stream Partitioning
|
||||
|
||||
Streaming everything through a single connection can create noisy event traffic—especially when mixing token-level text and tool telemetry. A clean architecture can **partition the stream** into two logical channels:
|
||||
|
||||
* **Main Assistant Stream:** carries incremental text deltas (`content` events, message tokens, etc.)
|
||||
* **Tool Call Stream:** carries system-level events (`tool_call`, `tool_result`, errors, and logs)
|
||||
|
||||
This can be implemented as separate **Server-Sent Event (SSE)** endpoints or separate NDJSON channels multiplexed within the same connection.
|
||||
|
||||
Partitioning provides better observability and reduces frontend coupling: you can display conversation flow independently of background tool actions, yet synchronize them via IDs or timestamps.
|
||||
|
||||
---
|
||||
|
||||
### H2 - 6. Production Validation and Observability
|
||||
|
||||
Before deployment, stress-test the entire flow with simulated long-running tools (e.g., artificial delays or heavy operations). During these tests:
|
||||
|
||||
* Verify that `"tool_call"` events appear as soon as the tool starts.
|
||||
* Ensure `"tool_result"` events arrive even when network latency or retries occur.
|
||||
* Confirm that JSON objects remain intact regardless of chunk size or proxy buffering.
|
||||
* Test browser-side resilience—if a user refreshes mid-stream, does the system resume cleanly or gracefully fail?
|
||||
|
||||
Integrate verbose logging at the stream layer to record:
|
||||
|
||||
* Timing between event types
|
||||
* Average duration of tool calls
|
||||
* Frequency of malformed JSON (useful for debugging agent misbehavior)
|
||||
|
||||
This telemetry will inform later performance optimizations and model-prompt adjustments.
|
||||
|
||||
---
|
||||
|
||||
## H1 - Technical Underpinnings
|
||||
|
||||
### H2 - Streaming Transport
|
||||
|
||||
The underlying transport is **HTTP chunked transfer encoding**. The Webhook node in n8n keeps the connection open, writing partial data with `res.write()` and closing it with `res.end()`. Each chunk arrives at the frontend as part of a continuous readable stream.
|
||||
|
||||
This design requires a decoder capable of handling partial UTF-8 boundaries, incomplete JSON, and asynchronous event emission.
|
||||
|
||||
### H2 - Event Representation
|
||||
|
||||
Most streaming ecosystems (OpenAI, Anthropic, Cohere, Vercel AI SDK) use event-based framing—either NDJSON (newline-delimited JSON) or SSE (Server-Sent Events). Both allow discrete events to be interpreted incrementally.
|
||||
|
||||
By framing your events this way, you gain interoperability with existing streaming libraries and avoid ambiguity between textual output and structured control messages.
|
||||
|
||||
### H2 - Workflow Integration
|
||||
|
||||
Within n8n:
|
||||
|
||||
* The **Webhook node** provides the live output channel.
|
||||
* The **AI Agent node** generates streaming data.
|
||||
* Custom **Code nodes** or **Function items** can intercept and re-emit synthetic events.
|
||||
|
||||
Because n8n allows direct access to the HTTP response object when using “Response mode: Streaming,” you can write custom control frames in exactly the same stream, ensuring chronological integrity between text and tool events.
|
||||
|
||||
### H2 - State Management and Error Handling
|
||||
|
||||
The capture layer functions as a **finite-state machine** managing transitions:
|
||||
|
||||
* `idle → capturing → emitting → idle`
|
||||
* With sub-states for quoted strings, escapes, and brace balancing.
|
||||
|
||||
When malformed JSON or premature stream closure occurs, the system should reset state, log the anomaly, and optionally emit a `type:"error"` event for client awareness. This ensures stability under unpredictable streaming conditions.
|
||||
|
||||
---
|
||||
|
||||
## H1 - Future Integration and Scaling
|
||||
|
||||
When native n8n support for tool streaming becomes available, your current architecture will need only minor adjustments:
|
||||
|
||||
* Replace the custom “before/after” emitters with n8n’s built-in event hooks.
|
||||
* Switch to official event types (`tool_call_start`, `tool_call_result`) without changing your frontend logic.
|
||||
* Optionally drop the JSON-capture layer if n8n begins emitting well-framed event data by default.
|
||||
|
||||
Long term, this foundation supports parallel streams, multi-agent collaboration, and event-driven orchestration—making Correspondents extensible for real-time AI systems that mix reasoning and automation.
|
||||
|
||||
---
|
||||
|
||||
## H1 - Summary
|
||||
|
||||
1. **Preserve your schema.** Flat JSON remains the most reliable format for deterministic streaming.
|
||||
2. **Reassemble fragmented JSON.** Introduce a lightweight capture layer to maintain structural integrity.
|
||||
3. **Expose tool events manually.** Emit start/result signals around tool executions for transparency.
|
||||
4. **Monitor native updates.** Track n8n’s release progress for built-in tool-call streaming.
|
||||
5. **Partition streams.** Keep text and system events logically distinct for clarity and reliability.
|
||||
6. **Validate continuously.** Use long-running simulations and logging to confirm real-time correctness.
|
||||
|
||||
This strategy balances robustness with forward compatibility—ensuring Correspondents can stream dynamic agent behavior today while seamlessly aligning with future n8n capabilities.
|
||||
|
||||
**Tags:** #streaming #n8n #architecture #toolcalls #Correspondents
|
||||
|
||||
184
docs/n8n-webhook-streaming.md
Normal file
184
docs/n8n-webhook-streaming.md
Normal file
@ -0,0 +1,184 @@
|
||||
# Full report on how streaming actually works with webhooks in n8n
|
||||
|
||||
(what is happening under the hood, why the versions matter, and how you wire this into your Correspondents stack)
|
||||
|
||||
I’m going to walk through this in a very plain, conversational way rather than repeating docs. The goal is to give you the “real mental model” of how n8n handles streaming so you can build a robust agent API for Correspondents.
|
||||
|
||||
---
|
||||
|
||||
# H1 - What “streaming” means inside n8n
|
||||
|
||||
n8n does not stream in the sense of WebSockets or Server Sent Events.
|
||||
It uses plain HTTP chunked transfer - basically the node writes multiple `res.write()` chunks to the webhook connection until the workflow ends, then does a final `res.end()`.
|
||||
|
||||
So your frontend - agents.nicholai.work - needs to be able to read the chunks as they come in. Libraries like fetch-with-streaming, readable streams, or SSE-like wrappers work fine.
|
||||
|
||||
There is no buffering on n8n’s side once streaming is enabled. Each node that supports streaming emits pieces of data as they are produced.
|
||||
|
||||
---
|
||||
|
||||
# H1 - Why version 1.105.2+ matters
|
||||
|
||||
Before ~1.105.x, the Webhook node hard-terminated the response early and the AI Agent node didn’t expose the streaming flag publicly.
|
||||
|
||||
After 1.105.2:
|
||||
|
||||
* The Webhook node gained a true “Streaming” response mode that keeps the HTTP response open.
|
||||
* The AI Agent node gained support for chunked output and a `stream: true` flag internally.
|
||||
* n8n’s runtime gained a proper `pushChunk` pipeline - meaning nodes can flush data without waiting for the workflow to finish.
|
||||
|
||||
Your Correspondents architecture depends on this new runtime. If you're under that version, the workflow waits until completion and dumps one JSON blob.
|
||||
|
||||
---
|
||||
|
||||
# H1 - The real mechanics: how the Webhook node streams
|
||||
|
||||
When you set the Webhook node to “Response mode: Streaming”, three things happen:
|
||||
|
||||
## H2 - 1. n8n tells Express not to auto-close the response
|
||||
|
||||
This stops the default behavior where a workflow finishes and n8n auto-sends the output.
|
||||
|
||||
## H2 - 2. The node switches into “manual response mode”
|
||||
|
||||
`res.write()` becomes available to downstream nodes.
|
||||
|
||||
## H2 - 3. The workflow execution channel is kept alive
|
||||
|
||||
n8n's internal worker uses a duplex stream so that downstream nodes can emit arbitrary numbers of chunks.
|
||||
|
||||
That is the entire magic. It’s simple once you know what's going on.
|
||||
|
||||
---
|
||||
|
||||
# H1 - How the AI Agent node streams
|
||||
|
||||
The AI Agent node is built on top of the new n8n LLM abstraction layer (which wraps provider SDKs like OpenAI, Anthropic, Mistral, Groq, etc).
|
||||
|
||||
When you enable streaming in the AI Agent node:
|
||||
|
||||
* The node uses the provider’s native streaming API
|
||||
* Each token or chunk triggers a callback
|
||||
* The callback uses `this.sendMessageToUI` for debugging and `this.pushOutput` for the webhook stream
|
||||
* The Webhook node emits each chunk to the client as a separate write
|
||||
|
||||
So the data goes like this:
|
||||
|
||||
Provider → AI Agent Node → n8n chunk buffer → Webhook → your client
|
||||
|
||||
Nothing sits in memory waiting for completion unless the model provider itself has that behavior.
|
||||
|
||||
---
|
||||
|
||||
# H1 - The correct wiring for your Correspondents architecture
|
||||
|
||||
Your workflow needs to be shaped like this:
|
||||
|
||||
Webhook (Streaming)
|
||||
→ Parse Request
|
||||
→ AI Agent (streaming enabled)
|
||||
→ (optional) transforms
|
||||
→ Webhook Respond (or not needed if streaming is active)
|
||||
|
||||
You **do not** use a "Webhook Respond" node in streaming mode.
|
||||
The Webhook node itself ends the connection when the workflow finishes.
|
||||
|
||||
So your workflow ends with the AI Agent node, or a final “completion” function, but no explicit response node.
|
||||
|
||||
---
|
||||
|
||||
# H1 - What your client must do
|
||||
|
||||
Since the n8n webhook responses are plain HTTP chunks, your client needs to read a **ReadableStream**.
|
||||
|
||||
Your frontend will look something like this (shortened for clarity):
|
||||
|
||||
```
|
||||
const response = await fetch(url, { method: "POST", body: payload });
|
||||
const reader = response.body.getReader();
|
||||
|
||||
while (true) {
|
||||
const { done, value } = await reader.read();
|
||||
if (done) break;
|
||||
const text = new TextDecoder().decode(value);
|
||||
// handle chunk...
|
||||
}
|
||||
```
|
||||
|
||||
That is literally all streaming requires on your side.
|
||||
|
||||
---
|
||||
|
||||
# H1 - Known pitfalls that bite real production workflows
|
||||
|
||||
## H2 - 1. Using the old AI nodes
|
||||
|
||||
If you created your workflow before 1.105.x, you need to delete and re-add:
|
||||
|
||||
* Webhook node
|
||||
* AI Agent node
|
||||
|
||||
n8n hard-caches node versions per-workflow.
|
||||
|
||||
## H2 - 2. Returning JSON inside a streaming workflow
|
||||
|
||||
You cannot stream and then return JSON at the end.
|
||||
Streaming means the connection ends when the workflow ends - no trailing payload allowed.
|
||||
|
||||
## H2 - 3. Host reverse-proxies sometimes buffer chunks
|
||||
|
||||
Cloudflare, Nginx, Traefik, Caddy can all buffer unless explicitly configured not to.
|
||||
n8n’s own Cloud-hosted version solves this for you, but self-host setups need:
|
||||
|
||||
`proxy_buffering off;`
|
||||
|
||||
or equivalent.
|
||||
|
||||
## H2 - 4. AI Agent streaming only works for supported providers
|
||||
|
||||
Anthropic, OpenAI, Groq, Mistral etc.
|
||||
If you use a provider that n8n wraps via HTTP only, streaming may be faked or disabled.
|
||||
|
||||
---
|
||||
|
||||
# H1 - How this ties directly into your Correspondents repo
|
||||
|
||||
Your architecture is:
|
||||
|
||||
agents.nicholai.work
|
||||
→ webhook trigger (stream)
|
||||
→ agent logic (custom)
|
||||
→ n8n AI Agent node (stream)
|
||||
→ stream back to client until agent finishes
|
||||
|
||||
This means you can implement:
|
||||
|
||||
* GPT style token streaming
|
||||
* Multi-agent streaming
|
||||
* Stream partial tool results
|
||||
* Stream logs or “thoughts” like OpenAI Logprobs / reasoning
|
||||
|
||||
As long as each chunk is sent as plain text, the client sees it instantly.
|
||||
|
||||
If you want to multiplex multiple channels (logs, events, tokens), you can prefix chunks:
|
||||
|
||||
```
|
||||
event:token Hello
|
||||
event:log Running step 1
|
||||
event:token world
|
||||
```
|
||||
|
||||
And your client router can handle it on your end.
|
||||
|
||||
---
|
||||
|
||||
# H1 - Final summary in normal English, no fluff
|
||||
|
||||
Streaming in n8n is just chunked HTTP responses.
|
||||
The Webhook node keeps the HTTP connection open.
|
||||
The AI Agent node emits tokens as they arrive from the model provider.
|
||||
Your client reads chunks.
|
||||
No magic beyond that.
|
||||
|
||||
This gives you a fully ChatGPT-like real time experience inside n8n workflows, including multi-agent setups like Correspondents.
|
||||
|
||||
@ -3,8 +3,8 @@
|
||||
"version": "0.1.0",
|
||||
"private": true,
|
||||
"scripts": {
|
||||
"dev": "next dev",
|
||||
"build": "next build",
|
||||
"dev": "node scripts/generate-morgan-prompt.js && next dev",
|
||||
"build": "node scripts/generate-morgan-prompt.js && next build",
|
||||
"start": "next start",
|
||||
"lint": "next lint",
|
||||
"test": "vitest",
|
||||
@ -14,6 +14,7 @@
|
||||
"dependencies": {
|
||||
"@hookform/resolvers": "^3.10.0",
|
||||
"@opennextjs/cloudflare": "^1.12.0",
|
||||
"@openrouter/ai-sdk-provider": "^1.2.3",
|
||||
"@radix-ui/react-accordion": "1.2.2",
|
||||
"@radix-ui/react-alert-dialog": "1.1.4",
|
||||
"@radix-ui/react-aspect-ratio": "1.1.1",
|
||||
@ -42,6 +43,7 @@
|
||||
"@radix-ui/react-toggle-group": "1.1.1",
|
||||
"@radix-ui/react-tooltip": "1.1.6",
|
||||
"@vercel/analytics": "1.3.1",
|
||||
"ai": "^5.0.93",
|
||||
"autoprefixer": "^10.4.20",
|
||||
"class-variance-authority": "^0.7.1",
|
||||
"clsx": "^2.1.1",
|
||||
@ -67,8 +69,9 @@
|
||||
"sonner": "^1.7.4",
|
||||
"tailwind-merge": "^3.3.1",
|
||||
"tailwindcss-animate": "^1.0.7",
|
||||
"uuid": "^13.0.0",
|
||||
"vaul": "^0.9.9",
|
||||
"zod": "3.25.67"
|
||||
"zod": "3.25.76"
|
||||
},
|
||||
"devDependencies": {
|
||||
"@tailwindcss/postcss": "^4.1.9",
|
||||
|
||||
159
pnpm-lock.yaml
generated
159
pnpm-lock.yaml
generated
@ -14,6 +14,9 @@ importers:
|
||||
'@opennextjs/cloudflare':
|
||||
specifier: ^1.12.0
|
||||
version: 1.12.0(wrangler@4.48.0)
|
||||
'@openrouter/ai-sdk-provider':
|
||||
specifier: ^1.2.3
|
||||
version: 1.2.3(ai@5.0.93(zod@3.25.76))(react-dom@19.1.0(react@19.1.0))(react@19.1.0)(zod@3.25.76)
|
||||
'@radix-ui/react-accordion':
|
||||
specifier: 1.2.2
|
||||
version: 1.2.2(@types/react-dom@18.3.7(@types/react@18.3.26))(@types/react@18.3.26)(react-dom@19.1.0(react@19.1.0))(react@19.1.0)
|
||||
@ -97,7 +100,10 @@ importers:
|
||||
version: 1.1.6(@types/react-dom@18.3.7(@types/react@18.3.26))(@types/react@18.3.26)(react-dom@19.1.0(react@19.1.0))(react@19.1.0)
|
||||
'@vercel/analytics':
|
||||
specifier: 1.3.1
|
||||
version: 1.3.1(next@15.5.4(@babel/core@7.28.5)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(react@19.1.0)
|
||||
version: 1.3.1(next@15.5.4(@babel/core@7.28.5)(@opentelemetry/api@1.9.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(react@19.1.0)
|
||||
ai:
|
||||
specifier: ^5.0.93
|
||||
version: 5.0.93(zod@3.25.76)
|
||||
autoprefixer:
|
||||
specifier: ^10.4.20
|
||||
version: 10.4.22(postcss@8.5.6)
|
||||
@ -124,7 +130,7 @@ importers:
|
||||
version: 12.23.24(react-dom@19.1.0(react@19.1.0))(react@19.1.0)
|
||||
geist:
|
||||
specifier: ^1.3.1
|
||||
version: 1.5.1(next@15.5.4(@babel/core@7.28.5)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))
|
||||
version: 1.5.1(next@15.5.4(@babel/core@7.28.5)(@opentelemetry/api@1.9.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))
|
||||
input-otp:
|
||||
specifier: 1.4.1
|
||||
version: 1.4.1(react-dom@19.1.0(react@19.1.0))(react@19.1.0)
|
||||
@ -133,7 +139,7 @@ importers:
|
||||
version: 0.454.0(react@19.1.0)
|
||||
next:
|
||||
specifier: 15.5.4
|
||||
version: 15.5.4(@babel/core@7.28.5)(react-dom@19.1.0(react@19.1.0))(react@19.1.0)
|
||||
version: 15.5.4(@babel/core@7.28.5)(@opentelemetry/api@1.9.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0)
|
||||
next-themes:
|
||||
specifier: ^0.4.6
|
||||
version: 0.4.6(react-dom@19.1.0(react@19.1.0))(react@19.1.0)
|
||||
@ -173,12 +179,15 @@ importers:
|
||||
tailwindcss-animate:
|
||||
specifier: ^1.0.7
|
||||
version: 1.0.7(tailwindcss@4.1.17)
|
||||
uuid:
|
||||
specifier: ^13.0.0
|
||||
version: 13.0.0
|
||||
vaul:
|
||||
specifier: ^0.9.9
|
||||
version: 0.9.9(@types/react-dom@18.3.7(@types/react@18.3.26))(@types/react@18.3.26)(react-dom@19.1.0(react@19.1.0))(react@19.1.0)
|
||||
zod:
|
||||
specifier: 3.25.67
|
||||
version: 3.25.67
|
||||
specifier: 3.25.76
|
||||
version: 3.25.76
|
||||
devDependencies:
|
||||
'@tailwindcss/postcss':
|
||||
specifier: ^4.1.9
|
||||
@ -240,6 +249,22 @@ packages:
|
||||
'@adobe/css-tools@4.4.4':
|
||||
resolution: {integrity: sha512-Elp+iwUx5rN5+Y8xLt5/GRoG20WGoDCQ/1Fb+1LiGtvwbDavuSk0jhD/eZdckHAuzcDzccnkv+rEjyWfRx18gg==}
|
||||
|
||||
'@ai-sdk/gateway@2.0.9':
|
||||
resolution: {integrity: sha512-E6x4h5CPPPJ0za1r5HsLtHbeI+Tp3H+YFtcH8G3dSSPFE6w+PZINzB4NxLZmg1QqSeA5HTP3ZEzzsohp0o2GEw==}
|
||||
engines: {node: '>=18'}
|
||||
peerDependencies:
|
||||
zod: ^3.25.76 || ^4.1.8
|
||||
|
||||
'@ai-sdk/provider-utils@3.0.17':
|
||||
resolution: {integrity: sha512-TR3Gs4I3Tym4Ll+EPdzRdvo/rc8Js6c4nVhFLuvGLX/Y4V9ZcQMa/HTiYsHEgmYrf1zVi6Q145UEZUfleOwOjw==}
|
||||
engines: {node: '>=18'}
|
||||
peerDependencies:
|
||||
zod: ^3.25.76 || ^4.1.8
|
||||
|
||||
'@ai-sdk/provider@2.0.0':
|
||||
resolution: {integrity: sha512-6o7Y2SeO9vFKB8lArHXehNuusnpddKPk7xqL7T2/b+OvXMRIXUO1rR4wcv1hAFUAT9avGZshty3Wlua/XA7TvA==}
|
||||
engines: {node: '>=18'}
|
||||
|
||||
'@alloc/quick-lru@5.2.0':
|
||||
resolution: {integrity: sha512-UrcABB+4bUrFABwbluTIBErXwvbsU/V7TZWfmbgJfbkwiBuziS9gxdODUyuiecfdGQ85jglMW6juS3+z5TsKLw==}
|
||||
engines: {node: '>=10'}
|
||||
@ -1391,6 +1416,31 @@ packages:
|
||||
peerDependencies:
|
||||
wrangler: ^4.38.0
|
||||
|
||||
'@openrouter/ai-sdk-provider@1.2.3':
|
||||
resolution: {integrity: sha512-a6Nc8dPRHakRH9966YJ/HZJhLOds7DuPTscNZDoAr+Aw+tEFUlacSJMvb/b3gukn74mgbuaJRji9YOn62ipfVg==}
|
||||
engines: {node: '>=18'}
|
||||
peerDependencies:
|
||||
ai: ^5.0.0
|
||||
zod: ^3.24.1 || ^v4
|
||||
|
||||
'@openrouter/sdk@0.1.11':
|
||||
resolution: {integrity: sha512-OuPc8qqidL/PUM8+9WgrOfSR9+b6rKIWiezGcUJ54iPTdh+Gye5Qjut6hrLWlOCMZE7Z853gN90r1ft4iChj7Q==}
|
||||
peerDependencies:
|
||||
'@tanstack/react-query': ^5
|
||||
react: ^18 || ^19
|
||||
react-dom: ^18 || ^19
|
||||
peerDependenciesMeta:
|
||||
'@tanstack/react-query':
|
||||
optional: true
|
||||
react:
|
||||
optional: true
|
||||
react-dom:
|
||||
optional: true
|
||||
|
||||
'@opentelemetry/api@1.9.0':
|
||||
resolution: {integrity: sha512-3giAOQvZiH5F9bMlMiv8+GSPMeqg0dbaeo58/0SlA9sxSqZhnUtxzX9/2FzyhS9sWQf5S0GJE0AKBrFqjpeYcg==}
|
||||
engines: {node: '>=8.0.0'}
|
||||
|
||||
'@poppinss/colors@4.1.5':
|
||||
resolution: {integrity: sha512-FvdDqtcRCtz6hThExcFOgW0cWX+xwSMWcRuQe5ZEb2m7cVQOAVZOIMt+/v9RxGiD9/OY16qJBXK4CVKWAPalBw==}
|
||||
|
||||
@ -2933,6 +2983,10 @@ packages:
|
||||
react:
|
||||
optional: true
|
||||
|
||||
'@vercel/oidc@3.0.3':
|
||||
resolution: {integrity: sha512-yNEQvPcVrK9sIe637+I0jD6leluPxzwJKx/Haw6F4H77CdDsszUn5V3o96LPziXkSNE2B83+Z3mjqGKBK/R6Gg==}
|
||||
engines: {node: '>= 20'}
|
||||
|
||||
'@vitejs/plugin-react@5.1.1':
|
||||
resolution: {integrity: sha512-WQfkSw0QbQ5aJ2CHYw23ZGkqnRwqKHD/KYsMeTkZzPT4Jcf0DcBxBtwMJxnu6E7oxw5+JC6ZAiePgh28uJ1HBA==}
|
||||
engines: {node: ^20.19.0 || >=22.12.0}
|
||||
@ -3007,6 +3061,12 @@ packages:
|
||||
resolution: {integrity: sha512-kja8j7PjmncONqaTsB8fQ+wE2mSU2DJ9D4XKoJ5PFWIdRMa6SLSN1ff4mOr4jCbfRSsxR4keIiySJU0N9T5hIQ==}
|
||||
engines: {node: '>= 8.0.0'}
|
||||
|
||||
ai@5.0.93:
|
||||
resolution: {integrity: sha512-9eGcu+1PJgPg4pRNV4L7tLjRR3wdJC9CXQoNMvtqvYNOLZHFCzjHtVIOr2SIkoJJeu2+sOy3hyiSuTmy2MA40g==}
|
||||
engines: {node: '>=18'}
|
||||
peerDependencies:
|
||||
zod: ^3.25.76 || ^4.1.8
|
||||
|
||||
ajv@6.12.6:
|
||||
resolution: {integrity: sha512-j3fVLgvTo527anyYyJOGTYJbG+vnnQYvE0m5mmkc1TK+nxAppkCLMIL0aZ4dblVCNoGShhm+kzE4ZUykBoMg4g==}
|
||||
|
||||
@ -3709,6 +3769,10 @@ packages:
|
||||
eventemitter3@4.0.7:
|
||||
resolution: {integrity: sha512-8guHBZCwKnFhYdHr2ysuRWErTwhoN2X8XELRlrRwpmfeY2jjuUN4taQMsULKUVo1K4DvZl+0pgfyoysHxvmvEw==}
|
||||
|
||||
eventsource-parser@3.0.6:
|
||||
resolution: {integrity: sha512-Vo1ab+QXPzZ4tCa8SwIHJFaSzy4R6SHf7BY79rFBDf0idraZWAkYrDjDj8uWaSm3S2TK+hJ7/t1CEmZ7jXw+pg==}
|
||||
engines: {node: '>=18.0.0'}
|
||||
|
||||
execa@5.1.1:
|
||||
resolution: {integrity: sha512-8uSpZZocAZRBAPIEINJj3Lo9HyGitllczc27Eh5YYojjMFMn8yHMDMaUHE2Jqfq05D/wucwI4JGURyXt1vchyg==}
|
||||
engines: {node: '>=10'}
|
||||
@ -4256,6 +4320,9 @@ packages:
|
||||
json-schema-traverse@0.4.1:
|
||||
resolution: {integrity: sha512-xbbCH5dCYU5T8LcEhhuh7HJ88HXuW3qsI3Y0zOZFKfZEHcpWiHU/Jxzk629Brsab/mMiHQti9wMP+845RPe3Vg==}
|
||||
|
||||
json-schema@0.4.0:
|
||||
resolution: {integrity: sha512-es94M3nTIfsEPisRafak+HDLfHXnKBhV3vU5eqPcS3flIWqcxJWgXHXiey3YrpaNsanY5ei1VoYEbOzijuq9BA==}
|
||||
|
||||
json-stable-stringify-without-jsonify@1.0.1:
|
||||
resolution: {integrity: sha512-Bdboy+l7tA3OGW6FjyFHWkP5LuByj1Tk33Ljyq0axyzdk9//JSi2u3fP1QSmd1KNwq6VOKYGlAu87CisVir6Pw==}
|
||||
|
||||
@ -5523,6 +5590,10 @@ packages:
|
||||
resolution: {integrity: sha512-pMZTvIkT1d+TFGvDOqodOclx0QWkkgi6Tdoa8gC8ffGAAqz9pzPTZWAybbsHHoED/ztMtkv/VoYTYyShUn81hA==}
|
||||
engines: {node: '>= 0.4.0'}
|
||||
|
||||
uuid@13.0.0:
|
||||
resolution: {integrity: sha512-XQegIaBTVUjSHliKqcnFqYypAd4S+WCYt5NIeRs6w/UAry7z8Y9j5ZwRRL4kzq9U3sD6v+85er9FvkEaBpji2w==}
|
||||
hasBin: true
|
||||
|
||||
uuid@9.0.1:
|
||||
resolution: {integrity: sha512-b+1eJOlsR9K8HJpow9Ok3fiWOWSIcIzXodvv0rQjVoOVNpWMpxf1wZNpt4y9h10odCNrqnYp1OBzRktckBe3sA==}
|
||||
hasBin: true
|
||||
@ -5785,8 +5856,8 @@ packages:
|
||||
zod@3.22.3:
|
||||
resolution: {integrity: sha512-EjIevzuJRiRPbVH4mGc8nApb/lVLKVpmUhAaR5R5doKGfAnGJ6Gr3CViAVjP+4FWSxCsybeWQdcgCtbX+7oZug==}
|
||||
|
||||
zod@3.25.67:
|
||||
resolution: {integrity: sha512-idA2YXwpCdqUSKRCACDE6ItZD9TZzy3OZMtpfLoh6oPR47lipysRrJfjzMqFxQ3uJuUPyUeWe1r9vLH33xO/Qw==}
|
||||
zod@3.25.76:
|
||||
resolution: {integrity: sha512-gzUt/qt81nXsFGKIFcC3YnfEAx5NkunCfnDlvuBSSFS02bcXu4Lmea0AFIUwbLWxWPx3d9p8S5QoaujKcNQxcQ==}
|
||||
|
||||
zwitch@2.0.4:
|
||||
resolution: {integrity: sha512-bXE4cR/kVZhKZX/RjPEflHaKVhUVl85noU3v6b8apfQEc1x4A+zBxjZ4lN8LqGd6WZ3dl98pY4o717VFmoPp+A==}
|
||||
@ -5797,6 +5868,24 @@ snapshots:
|
||||
|
||||
'@adobe/css-tools@4.4.4': {}
|
||||
|
||||
'@ai-sdk/gateway@2.0.9(zod@3.25.76)':
|
||||
dependencies:
|
||||
'@ai-sdk/provider': 2.0.0
|
||||
'@ai-sdk/provider-utils': 3.0.17(zod@3.25.76)
|
||||
'@vercel/oidc': 3.0.3
|
||||
zod: 3.25.76
|
||||
|
||||
'@ai-sdk/provider-utils@3.0.17(zod@3.25.76)':
|
||||
dependencies:
|
||||
'@ai-sdk/provider': 2.0.0
|
||||
'@standard-schema/spec': 1.0.0
|
||||
eventsource-parser: 3.0.6
|
||||
zod: 3.25.76
|
||||
|
||||
'@ai-sdk/provider@2.0.0':
|
||||
dependencies:
|
||||
json-schema: 0.4.0
|
||||
|
||||
'@alloc/quick-lru@5.2.0': {}
|
||||
|
||||
'@asamuzakjp/css-color@4.0.5':
|
||||
@ -7505,6 +7594,25 @@ snapshots:
|
||||
- encoding
|
||||
- supports-color
|
||||
|
||||
'@openrouter/ai-sdk-provider@1.2.3(ai@5.0.93(zod@3.25.76))(react-dom@19.1.0(react@19.1.0))(react@19.1.0)(zod@3.25.76)':
|
||||
dependencies:
|
||||
'@openrouter/sdk': 0.1.11(react-dom@19.1.0(react@19.1.0))(react@19.1.0)
|
||||
ai: 5.0.93(zod@3.25.76)
|
||||
zod: 3.25.76
|
||||
transitivePeerDependencies:
|
||||
- '@tanstack/react-query'
|
||||
- react
|
||||
- react-dom
|
||||
|
||||
'@openrouter/sdk@0.1.11(react-dom@19.1.0(react@19.1.0))(react@19.1.0)':
|
||||
dependencies:
|
||||
zod: 3.25.76
|
||||
optionalDependencies:
|
||||
react: 19.1.0
|
||||
react-dom: 19.1.0(react@19.1.0)
|
||||
|
||||
'@opentelemetry/api@1.9.0': {}
|
||||
|
||||
'@poppinss/colors@4.1.5':
|
||||
dependencies:
|
||||
kleur: 4.1.5
|
||||
@ -9231,13 +9339,15 @@ snapshots:
|
||||
'@unrs/resolver-binding-win32-x64-msvc@1.11.1':
|
||||
optional: true
|
||||
|
||||
'@vercel/analytics@1.3.1(next@15.5.4(@babel/core@7.28.5)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(react@19.1.0)':
|
||||
'@vercel/analytics@1.3.1(next@15.5.4(@babel/core@7.28.5)(@opentelemetry/api@1.9.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(react@19.1.0)':
|
||||
dependencies:
|
||||
server-only: 0.0.1
|
||||
optionalDependencies:
|
||||
next: 15.5.4(@babel/core@7.28.5)(react-dom@19.1.0(react@19.1.0))(react@19.1.0)
|
||||
next: 15.5.4(@babel/core@7.28.5)(@opentelemetry/api@1.9.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0)
|
||||
react: 19.1.0
|
||||
|
||||
'@vercel/oidc@3.0.3': {}
|
||||
|
||||
'@vitejs/plugin-react@5.1.1(vite@7.2.2(@types/node@22.19.1)(jiti@2.6.1)(lightningcss@1.30.2)(terser@5.16.9)(yaml@2.8.1))':
|
||||
dependencies:
|
||||
'@babel/core': 7.28.5
|
||||
@ -9316,6 +9426,14 @@ snapshots:
|
||||
dependencies:
|
||||
humanize-ms: 1.2.1
|
||||
|
||||
ai@5.0.93(zod@3.25.76):
|
||||
dependencies:
|
||||
'@ai-sdk/gateway': 2.0.9(zod@3.25.76)
|
||||
'@ai-sdk/provider': 2.0.0
|
||||
'@ai-sdk/provider-utils': 3.0.17(zod@3.25.76)
|
||||
'@opentelemetry/api': 1.9.0
|
||||
zod: 3.25.76
|
||||
|
||||
ajv@6.12.6:
|
||||
dependencies:
|
||||
fast-deep-equal: 3.1.3
|
||||
@ -10072,8 +10190,8 @@ snapshots:
|
||||
'@babel/parser': 7.28.5
|
||||
eslint: 9.39.1(jiti@2.6.1)
|
||||
hermes-parser: 0.25.1
|
||||
zod: 3.25.67
|
||||
zod-validation-error: 4.0.2(zod@3.25.67)
|
||||
zod: 3.25.76
|
||||
zod-validation-error: 4.0.2(zod@3.25.76)
|
||||
transitivePeerDependencies:
|
||||
- supports-color
|
||||
|
||||
@ -10179,6 +10297,8 @@ snapshots:
|
||||
|
||||
eventemitter3@4.0.7: {}
|
||||
|
||||
eventsource-parser@3.0.6: {}
|
||||
|
||||
execa@5.1.1:
|
||||
dependencies:
|
||||
cross-spawn: 7.0.6
|
||||
@ -10362,9 +10482,9 @@ snapshots:
|
||||
|
||||
functions-have-names@1.2.3: {}
|
||||
|
||||
geist@1.5.1(next@15.5.4(@babel/core@7.28.5)(react-dom@19.1.0(react@19.1.0))(react@19.1.0)):
|
||||
geist@1.5.1(next@15.5.4(@babel/core@7.28.5)(@opentelemetry/api@1.9.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0)):
|
||||
dependencies:
|
||||
next: 15.5.4(@babel/core@7.28.5)(react-dom@19.1.0(react@19.1.0))(react@19.1.0)
|
||||
next: 15.5.4(@babel/core@7.28.5)(@opentelemetry/api@1.9.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0)
|
||||
|
||||
generator-function@2.0.1: {}
|
||||
|
||||
@ -10785,6 +10905,8 @@ snapshots:
|
||||
|
||||
json-schema-traverse@0.4.1: {}
|
||||
|
||||
json-schema@0.4.0: {}
|
||||
|
||||
json-stable-stringify-without-jsonify@1.0.1: {}
|
||||
|
||||
json5@1.0.2:
|
||||
@ -11358,7 +11480,7 @@ snapshots:
|
||||
react: 19.1.0
|
||||
react-dom: 19.1.0(react@19.1.0)
|
||||
|
||||
next@15.5.4(@babel/core@7.28.5)(react-dom@19.1.0(react@19.1.0))(react@19.1.0):
|
||||
next@15.5.4(@babel/core@7.28.5)(@opentelemetry/api@1.9.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0):
|
||||
dependencies:
|
||||
'@next/env': 15.5.4
|
||||
'@swc/helpers': 0.5.15
|
||||
@ -11376,6 +11498,7 @@ snapshots:
|
||||
'@next/swc-linux-x64-musl': 15.5.4
|
||||
'@next/swc-win32-arm64-msvc': 15.5.4
|
||||
'@next/swc-win32-x64-msvc': 15.5.4
|
||||
'@opentelemetry/api': 1.9.0
|
||||
sharp: 0.34.5
|
||||
transitivePeerDependencies:
|
||||
- '@babel/core'
|
||||
@ -12410,6 +12533,8 @@ snapshots:
|
||||
|
||||
utils-merge@1.0.1: {}
|
||||
|
||||
uuid@13.0.0: {}
|
||||
|
||||
uuid@9.0.1: {}
|
||||
|
||||
vary@1.1.2: {}
|
||||
@ -12672,12 +12797,12 @@ snapshots:
|
||||
cookie: 1.0.2
|
||||
youch-core: 0.3.3
|
||||
|
||||
zod-validation-error@4.0.2(zod@3.25.67):
|
||||
zod-validation-error@4.0.2(zod@3.25.76):
|
||||
dependencies:
|
||||
zod: 3.25.67
|
||||
zod: 3.25.76
|
||||
|
||||
zod@3.22.3: {}
|
||||
|
||||
zod@3.25.67: {}
|
||||
zod@3.25.76: {}
|
||||
|
||||
zwitch@2.0.4: {}
|
||||
|
||||
43
scripts/generate-morgan-prompt.js
Normal file
43
scripts/generate-morgan-prompt.js
Normal file
@ -0,0 +1,43 @@
|
||||
#!/usr/bin/env node
|
||||
|
||||
/**
|
||||
* Build-time script to generate TypeScript file with Morgan's system prompt
|
||||
* This embeds the markdown file as a constant for Cloudflare Worker deployment
|
||||
*/
|
||||
|
||||
const fs = require('fs')
|
||||
const path = require('path')
|
||||
|
||||
const sourceFile = path.join(__dirname, '..', '.fortura-core', 'web-agents', 'agent-architect-web.md')
|
||||
const outputFile = path.join(__dirname, '..', 'src', 'lib', 'agents', 'morgan-system-prompt.ts')
|
||||
|
||||
try {
|
||||
// Read the source markdown file
|
||||
const content = fs.readFileSync(sourceFile, 'utf-8')
|
||||
|
||||
// Escape backticks and other special characters for TypeScript string
|
||||
const escaped = content
|
||||
.replace(/\\/g, '\\\\') // Escape backslashes first
|
||||
.replace(/`/g, '\\`') // Escape backticks
|
||||
.replace(/\$/g, '\\$') // Escape dollar signs (for template literals)
|
||||
|
||||
// Generate TypeScript file
|
||||
const tsContent = `/**
|
||||
* Morgan's System Prompt
|
||||
* Generated at build time from .fortura-core/web-agents/agent-architect-web.md
|
||||
* This is embedded here for Cloudflare Worker deployment compatibility
|
||||
*/
|
||||
|
||||
export const MORGAN_SYSTEM_PROMPT = \`${escaped}\`
|
||||
`
|
||||
|
||||
// Write the output file
|
||||
fs.writeFileSync(outputFile, tsContent, 'utf-8')
|
||||
|
||||
console.log(`✓ Generated Morgan system prompt at ${outputFile}`)
|
||||
console.log(` Size: ${(tsContent.length / 1024).toFixed(1)} KB`)
|
||||
process.exit(0)
|
||||
} catch (error) {
|
||||
console.error('✗ Failed to generate Morgan system prompt:', error.message)
|
||||
process.exit(1)
|
||||
}
|
||||
@ -1,66 +1,40 @@
|
||||
import { type NextRequest, NextResponse } from "next/server"
|
||||
import type { Agent, AgentsResponse } from "@/lib/types"
|
||||
import { NextRequest, NextResponse } from 'next/server'
|
||||
import type { Agent, AgentsResponse } from '@/lib/types'
|
||||
import { getAllAvailableAgents } from '@/lib/agents/factory'
|
||||
|
||||
/**
|
||||
* GET /api/agents
|
||||
* Returns list of available agents configured via environment variables
|
||||
* Returns list of available agents from code-based definitions
|
||||
*
|
||||
* Expected environment variables format:
|
||||
* - AGENT_1_URL, AGENT_1_NAME, AGENT_1_DESCRIPTION
|
||||
* - AGENT_2_URL, AGENT_2_NAME, AGENT_2_DESCRIPTION
|
||||
* - etc.
|
||||
* Agents are now defined in src/lib/agents/definitions.ts
|
||||
* instead of environment variables
|
||||
*/
|
||||
export async function GET(request: NextRequest): Promise<NextResponse<AgentsResponse>> {
|
||||
try {
|
||||
const agents: Agent[] = []
|
||||
// Get standard agents from definitions
|
||||
const agentDefs = getAllAvailableAgents()
|
||||
|
||||
// Parse agent configurations from environment variables
|
||||
// Look for AGENT_N_URL, AGENT_N_NAME, AGENT_N_DESCRIPTION patterns
|
||||
let agentIndex = 1
|
||||
|
||||
while (true) {
|
||||
const urlKey = `AGENT_${agentIndex}_URL`
|
||||
const nameKey = `AGENT_${agentIndex}_NAME`
|
||||
const descriptionKey = `AGENT_${agentIndex}_DESCRIPTION`
|
||||
|
||||
const webhookUrl = process.env[urlKey]
|
||||
const name = process.env[nameKey]
|
||||
const description = process.env[descriptionKey]
|
||||
|
||||
// Stop if we don't find a URL for this index
|
||||
if (!webhookUrl) {
|
||||
break
|
||||
}
|
||||
|
||||
// Require at least URL and name
|
||||
if (!name) {
|
||||
console.warn(`[agents] Agent ${agentIndex} missing name, skipping`)
|
||||
agentIndex++
|
||||
continue
|
||||
}
|
||||
|
||||
agents.push({
|
||||
id: `agent-${agentIndex}`,
|
||||
name,
|
||||
description: description || "",
|
||||
webhookUrl,
|
||||
})
|
||||
|
||||
agentIndex++
|
||||
}
|
||||
// Convert to Agent format (client-side response)
|
||||
const agents: Agent[] = agentDefs.map((agent) => ({
|
||||
id: agent.id,
|
||||
name: agent.name,
|
||||
description: agent.description,
|
||||
// Note: webhookUrl is no longer used with Vercel AI SDK
|
||||
// All requests go through the unified /api/chat endpoint
|
||||
}))
|
||||
|
||||
if (agents.length === 0) {
|
||||
console.warn("[agents] No agents configured in environment variables")
|
||||
console.warn('[agents] No agents configured')
|
||||
}
|
||||
|
||||
console.log(`[agents] Loaded ${agents.length} agents`)
|
||||
|
||||
return NextResponse.json({ agents })
|
||||
} catch (error) {
|
||||
console.error("[agents] Error loading agents:", error)
|
||||
console.error('[agents] Error loading agents:', error)
|
||||
return NextResponse.json(
|
||||
{ agents: [], error: "Failed to load agents" },
|
||||
{ status: 500 },
|
||||
{ agents: [], error: 'Failed to load agents' },
|
||||
{ status: 500 }
|
||||
)
|
||||
}
|
||||
}
|
||||
|
||||
@ -1,348 +1,155 @@
|
||||
import { type NextRequest, NextResponse } from "next/server"
|
||||
import type { ChatRequest, ChatResponse } from "@/lib/types"
|
||||
import { getFlags } from "@/lib/flags"
|
||||
'use server'
|
||||
|
||||
import { streamText } from 'ai'
|
||||
import { NextRequest, NextResponse } from 'next/server'
|
||||
import type { ChatRequest } from '@/lib/types'
|
||||
import { getConfiguredModel } from '@/lib/openrouter'
|
||||
import { getAgentDefinition } from '@/lib/agents/factory'
|
||||
import { getFlags } from '@/lib/flags'
|
||||
|
||||
/**
|
||||
* Get webhook URL for a specific agent from environment variables
|
||||
* Format: AGENT_{agentIndex}_URL or custom agent handler
|
||||
* POST /api/chat
|
||||
* Stream chat responses from the selected agent
|
||||
*
|
||||
* Request body:
|
||||
* - message: User message
|
||||
* - agentId: Selected agent (agent-1, agent-2, or custom-{uuid})
|
||||
* - sessionId: Session ID for conversation tracking
|
||||
* - timestamp: Request timestamp
|
||||
* - images?: Base64 encoded images (optional)
|
||||
* - systemPrompt?: For custom agents, the agent's system prompt
|
||||
*
|
||||
* Response:
|
||||
* - Server-Sent Events (SSE) stream with text and tool calls
|
||||
*/
|
||||
function getAgentWebhookUrl(agentId: string): string | null {
|
||||
// Check if this is a custom agent (format: "custom-{id}")
|
||||
if (agentId.startsWith("custom-")) {
|
||||
// Custom agents use a dedicated webhook if configured
|
||||
const customWebhook = process.env.CUSTOM_AGENT_WEBHOOK
|
||||
if (customWebhook) {
|
||||
return customWebhook
|
||||
}
|
||||
console.error("[chat] No CUSTOM_AGENT_WEBHOOK configured for custom agents")
|
||||
return null
|
||||
}
|
||||
|
||||
// Extract agent index from agentId (format: "agent-1", "agent-2", etc.)
|
||||
const match = agentId.match(/agent-(\d+)/)
|
||||
if (!match) {
|
||||
console.error("[chat] Invalid agentId format:", agentId)
|
||||
return null
|
||||
}
|
||||
|
||||
const agentIndex = match[1]
|
||||
const urlKey = `AGENT_${agentIndex}_URL`
|
||||
const webhookUrl = process.env[urlKey]
|
||||
|
||||
if (!webhookUrl) {
|
||||
console.error(`[chat] No webhook URL configured for ${urlKey}`)
|
||||
return null
|
||||
}
|
||||
|
||||
return webhookUrl
|
||||
}
|
||||
|
||||
// Helper function to convert diff tool call to markdown format
|
||||
function convertToDiffTool(args: any, diffToolEnabled: boolean): string {
|
||||
export async function POST(request: NextRequest) {
|
||||
try {
|
||||
const { oldCode, newCode, title, language } = args
|
||||
|
||||
if (!oldCode || !newCode) {
|
||||
return "Error: Missing oldCode or newCode in diff tool call"
|
||||
// Parse request body
|
||||
const body = (await request.json()) as Partial<ChatRequest> & {
|
||||
systemPrompt?: string
|
||||
}
|
||||
|
||||
// If diff tool is disabled, return as plain code blocks
|
||||
if (!diffToolEnabled) {
|
||||
const titleText = title || "Code Changes"
|
||||
return `### ${titleText}\n\n**Before:**\n\`\`\`${language || 'text'}\n${oldCode}\n\`\`\`\n\n**After:**\n\`\`\`${language || 'text'}\n${newCode}\n\`\`\``
|
||||
}
|
||||
|
||||
const diffToolCall = {
|
||||
oldCode: String(oldCode).replace(/\n/g, '\\n'),
|
||||
newCode: String(newCode).replace(/\n/g, '\\n'),
|
||||
title: title || "Code Changes",
|
||||
language: language || "text"
|
||||
}
|
||||
|
||||
return `\`\`\`diff-tool\n${JSON.stringify(diffToolCall, null, 2)}\n\`\`\``
|
||||
} catch (error) {
|
||||
console.error("[v0] Error converting diff tool:", error)
|
||||
return "Error: Failed to process diff tool call"
|
||||
}
|
||||
}
|
||||
|
||||
export async function POST(request: NextRequest): Promise<NextResponse<ChatResponse>> {
|
||||
try {
|
||||
const body = await request.json()
|
||||
if (typeof body !== "object" || body === null) {
|
||||
return NextResponse.json({ error: "Invalid request body" }, { status: 400 })
|
||||
}
|
||||
|
||||
const { message, timestamp, sessionId, agentId, images, systemPrompt } = body as ChatRequest & { systemPrompt?: string }
|
||||
|
||||
// Get feature flags
|
||||
const flags = getFlags()
|
||||
const { message, agentId, sessionId, timestamp, images, systemPrompt } = body
|
||||
|
||||
// Validate required fields
|
||||
if (!message || typeof message !== "string") {
|
||||
return NextResponse.json({ error: "Message is required" }, { status: 400 })
|
||||
}
|
||||
|
||||
if (!agentId || typeof agentId !== "string") {
|
||||
return NextResponse.json({ error: "Agent ID is required" }, { status: 400 })
|
||||
}
|
||||
|
||||
// Validate systemPrompt for custom agents
|
||||
if (agentId.startsWith("custom-") && !systemPrompt) {
|
||||
if (!message) {
|
||||
return NextResponse.json(
|
||||
{ error: "systemPrompt is required for custom agents" },
|
||||
{ error: 'Message is required' },
|
||||
{ status: 400 }
|
||||
)
|
||||
}
|
||||
|
||||
// Check if image uploads are enabled
|
||||
if (!agentId) {
|
||||
return NextResponse.json(
|
||||
{ error: 'Agent ID is required' },
|
||||
{ status: 400 }
|
||||
)
|
||||
}
|
||||
|
||||
// Check feature flags
|
||||
const flags = getFlags()
|
||||
if (images && images.length > 0 && !flags.IMAGE_UPLOADS_ENABLED) {
|
||||
return NextResponse.json(
|
||||
{
|
||||
error: "Image uploads are currently disabled",
|
||||
hint: "Contact your administrator to enable the IMAGE_UPLOADS_ENABLED flag"
|
||||
error: 'Image uploads are not enabled',
|
||||
hint: 'Contact administrator to enable this feature',
|
||||
},
|
||||
{ status: 403 }
|
||||
)
|
||||
}
|
||||
|
||||
// Get webhook URL for the selected agent
|
||||
const webhookUrl = getAgentWebhookUrl(agentId)
|
||||
if (!webhookUrl) {
|
||||
return NextResponse.json(
|
||||
{ error: `Agent ${agentId} is not properly configured` },
|
||||
{ status: 400 },
|
||||
)
|
||||
}
|
||||
// Log request
|
||||
console.log(`[chat] Agent: ${agentId}, Session: ${sessionId}, Message length: ${message.length}`)
|
||||
|
||||
console.log("[chat] Sending to webhook:", { agentId, message, timestamp, sessionId })
|
||||
|
||||
const webhookPayload: any = {
|
||||
message,
|
||||
timestamp,
|
||||
sessionId,
|
||||
agentId,
|
||||
images: images && images.length > 0 ? images : undefined,
|
||||
}
|
||||
|
||||
// Include systemPrompt for custom agents
|
||||
if (systemPrompt) {
|
||||
webhookPayload.systemPrompt = systemPrompt
|
||||
}
|
||||
|
||||
const response = await fetch(webhookUrl, {
|
||||
method: "POST",
|
||||
headers: {
|
||||
"Content-Type": "application/json",
|
||||
},
|
||||
body: JSON.stringify(webhookPayload),
|
||||
// Load agent definition
|
||||
const agent = await getAgentDefinition(agentId, {
|
||||
systemPrompt: systemPrompt || '',
|
||||
tools: undefined, // Tools come from agent definition
|
||||
})
|
||||
|
||||
console.log("[v0] Webhook response status:", response.status)
|
||||
// Build message array with context
|
||||
const messageContent = buildMessageContent(message, images)
|
||||
const messages: Parameters<typeof streamText>[0]['messages'] = [
|
||||
{
|
||||
role: 'user',
|
||||
content: messageContent as any,
|
||||
},
|
||||
]
|
||||
|
||||
const responseText = await response.text()
|
||||
console.log("[v0] Webhook response body (first 200 chars):", responseText.substring(0, 200))
|
||||
// Get configured model
|
||||
const model = getConfiguredModel()
|
||||
|
||||
if (!response.ok) {
|
||||
// Try to parse as JSON if possible, otherwise use text
|
||||
let errorData
|
||||
try {
|
||||
errorData = responseText ? JSON.parse(responseText) : {}
|
||||
} catch {
|
||||
errorData = { message: responseText || "Unknown error" }
|
||||
}
|
||||
|
||||
console.error("[v0] Webhook error:", errorData)
|
||||
// Stream response from agent
|
||||
const result = streamText({
|
||||
model,
|
||||
system: agent.systemPrompt,
|
||||
tools: agent.tools || {},
|
||||
messages,
|
||||
temperature: agent.temperature,
|
||||
// Note: maxTokens is not used in streamText - it uses maxRetries, retry logic, etc
|
||||
onFinish: (event) => {
|
||||
console.log(`[chat] Response completed for agent ${agentId}`)
|
||||
},
|
||||
})
|
||||
|
||||
// Return as text stream response (Server-Sent Events format)
|
||||
return result.toTextStreamResponse()
|
||||
} catch (error) {
|
||||
console.error('[chat] Error:', error)
|
||||
const message = error instanceof Error ? error.message : 'Unknown error'
|
||||
return NextResponse.json(
|
||||
{
|
||||
error: errorData.message || "Failed to communicate with webhook",
|
||||
hint: errorData.hint,
|
||||
code: errorData.code,
|
||||
error: 'Failed to process message',
|
||||
hint: message,
|
||||
},
|
||||
{ status: response.status },
|
||||
{ status: 500 }
|
||||
)
|
||||
}
|
||||
|
||||
if (!responseText) {
|
||||
console.log("[v0] Empty response from webhook")
|
||||
return NextResponse.json({
|
||||
response:
|
||||
"The webhook received your message but didn't return a response. Please ensure your n8n workflow includes a 'Respond to Webhook' node that returns data.",
|
||||
hint: "Add a 'Respond to Webhook' node in your n8n workflow to send responses back to the chat.",
|
||||
})
|
||||
}
|
||||
|
||||
try {
|
||||
// First, check if the response contains a tool call in a markdown code block
|
||||
// This handles cases where n8n wraps the tool call in markdown
|
||||
const toolCallMatch = responseText.match(/```json\s*\n\s*(\{[\s\S]*?"type"\s*:\s*"tool_call"[\s\S]*?\})\s*\n\s*```/)
|
||||
if (toolCallMatch) {
|
||||
try {
|
||||
const toolCallJson = JSON.parse(toolCallMatch[1])
|
||||
if (toolCallJson.type === "tool_call" && toolCallJson.name === "create_agent_package") {
|
||||
console.log("[v0] Extracted tool call from markdown code block")
|
||||
return NextResponse.json({
|
||||
response: "",
|
||||
toolCall: toolCallJson
|
||||
})
|
||||
}
|
||||
} catch (error) {
|
||||
console.error("[v0] Failed to parse tool call from markdown:", error)
|
||||
}
|
||||
}
|
||||
|
||||
// Split response by newlines to get individual JSON objects
|
||||
const lines = responseText.trim().split("\n")
|
||||
const chunks: string[] = []
|
||||
|
||||
for (const line of lines) {
|
||||
if (!line.trim()) continue
|
||||
|
||||
try {
|
||||
const chunk = JSON.parse(line)
|
||||
|
||||
// Extract content from "item" type chunks
|
||||
if (chunk.type === "item" && chunk.content) {
|
||||
chunks.push(chunk.content)
|
||||
}
|
||||
|
||||
// Handle diff tool calls
|
||||
if (chunk.type === "tool_call" && chunk.name === "show_diff") {
|
||||
const diffTool = convertToDiffTool(chunk.args, flags.DIFF_TOOL_ENABLED)
|
||||
chunks.push(diffTool)
|
||||
}
|
||||
|
||||
// Handle agent package tool calls - forward as-is to client
|
||||
if (chunk.type === "tool_call" && chunk.name === "create_agent_package") {
|
||||
// Return the tool call directly so the client can handle it
|
||||
return NextResponse.json({
|
||||
response: "",
|
||||
toolCall: chunk
|
||||
})
|
||||
}
|
||||
} catch {
|
||||
console.log("[v0] Failed to parse line:", line)
|
||||
}
|
||||
}
|
||||
|
||||
// Combine all chunks into a single message
|
||||
if (chunks.length > 0) {
|
||||
const fullMessage = chunks.join("")
|
||||
console.log("[v0] Combined message from", chunks.length, "chunks")
|
||||
return NextResponse.json({ response: fullMessage })
|
||||
}
|
||||
|
||||
// If no chunks found, try parsing as regular JSON
|
||||
const data = JSON.parse(responseText)
|
||||
console.log("[v0] Parsed webhook data:", data)
|
||||
|
||||
// Handle n8n Code node output format: { output: { messageType: "...", content: "..." } }
|
||||
// Can be wrapped in array [{ output: {...} }] or direct { output: {...} }
|
||||
let parsedOutput = null
|
||||
|
||||
if (Array.isArray(data) && data.length > 0 && data[0].output) {
|
||||
parsedOutput = data[0].output
|
||||
} else if (data.output) {
|
||||
parsedOutput = data.output
|
||||
}
|
||||
|
||||
if (parsedOutput) {
|
||||
console.log("[v0] parsedOutput messageType:", parsedOutput.messageType)
|
||||
|
||||
if (parsedOutput?.messageType === "regular_message" && parsedOutput.content) {
|
||||
console.log("[v0] Code node output: regular message")
|
||||
return NextResponse.json({
|
||||
response: parsedOutput.content
|
||||
})
|
||||
}
|
||||
|
||||
if (parsedOutput?.messageType === "tool_call") {
|
||||
console.log("[v0] Code node output: tool call detected!")
|
||||
console.log("[v0] toolCall object:", parsedOutput.toolCall)
|
||||
// Tool calls have both content (narration) and toolCall (the actual data)
|
||||
const responseData = {
|
||||
response: parsedOutput.content || "",
|
||||
toolCall: parsedOutput.toolCall
|
||||
}
|
||||
console.log("[v0] Returning tool call response:", responseData)
|
||||
return NextResponse.json(responseData)
|
||||
}
|
||||
|
||||
console.log("[v0] parsedOutput exists but no messageType match")
|
||||
}
|
||||
|
||||
// Check if this is a diff tool call
|
||||
if (data.type === "tool_call" && data.name === "show_diff") {
|
||||
const diffTool = convertToDiffTool(data.args, flags.DIFF_TOOL_ENABLED)
|
||||
return NextResponse.json({ response: diffTool })
|
||||
}
|
||||
|
||||
// Check if this is an agent package tool call
|
||||
if (data.type === "tool_call" && data.name === "create_agent_package") {
|
||||
return NextResponse.json({
|
||||
response: "",
|
||||
toolCall: data
|
||||
})
|
||||
}
|
||||
|
||||
// Check if the response fields contain a markdown-wrapped OR plain JSON tool call
|
||||
const responseFields = [data.output, data.response, data.message, data.text].filter(Boolean)
|
||||
for (const field of responseFields) {
|
||||
if (typeof field === 'string') {
|
||||
// Try markdown-wrapped first
|
||||
let nestedToolCallMatch = field.match(/```json\s*\n\s*(\{[\s\S]*?"type"\s*:\s*"tool_call"[\s\S]*?\})\s*\n\s*```/)
|
||||
|
||||
// If no markdown wrapper, try plain JSON (with or without escape sequences)
|
||||
if (!nestedToolCallMatch) {
|
||||
// Match JSON object with "type": "tool_call" - handle both escaped and unescaped newlines
|
||||
const plainJsonMatch = field.match(/(\{[\s\S]*?"type"\s*:\s*"tool_call"[\s\S]*?\n\s*\})/)
|
||||
if (plainJsonMatch) {
|
||||
nestedToolCallMatch = plainJsonMatch
|
||||
}
|
||||
}
|
||||
|
||||
if (nestedToolCallMatch) {
|
||||
try {
|
||||
// Clean up the matched string - replace \n with actual newlines if needed
|
||||
let jsonString = nestedToolCallMatch[1]
|
||||
const toolCallJson = JSON.parse(jsonString)
|
||||
if (toolCallJson.type === "tool_call" && toolCallJson.name === "create_agent_package") {
|
||||
console.log("[v0] Extracted tool call from response field (plain or markdown)")
|
||||
return NextResponse.json({
|
||||
response: "",
|
||||
toolCall: toolCallJson
|
||||
})
|
||||
}
|
||||
} catch (error) {
|
||||
console.error("[v0] Failed to parse nested tool call:", error)
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// Extract the response from various possible fields
|
||||
let responseMessage = data.response || data.message || data.output || data.text
|
||||
|
||||
// If the response is an object, try to extract from nested fields
|
||||
if (typeof responseMessage === "object") {
|
||||
responseMessage =
|
||||
responseMessage.response || responseMessage.message || responseMessage.output || responseMessage.text
|
||||
}
|
||||
|
||||
// If still no message found, stringify the entire response
|
||||
if (!responseMessage) {
|
||||
responseMessage = JSON.stringify(data)
|
||||
}
|
||||
|
||||
return NextResponse.json({ response: responseMessage })
|
||||
} catch {
|
||||
console.log("[v0] Response is not JSON, returning as text")
|
||||
// If not JSON, return the text as the response
|
||||
return NextResponse.json({ response: responseText })
|
||||
}
|
||||
} catch (error) {
|
||||
console.error("[v0] API route error:", error)
|
||||
return NextResponse.json({ error: "Internal server error" }, { status: 500 })
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Build message content with text and images
|
||||
* Supports both text-only and multimodal messages
|
||||
*/
|
||||
function buildMessageContent(
|
||||
text: string,
|
||||
images?: string[]
|
||||
): string | Array<{ type: 'text' | 'image'; text?: string; image?: string; mimeType?: string }> {
|
||||
// Text only
|
||||
if (!images || images.length === 0) {
|
||||
return text
|
||||
}
|
||||
|
||||
// Multimodal message with images
|
||||
const content: Array<{ type: 'text' | 'image'; text?: string; image?: string; mimeType?: string }> = [
|
||||
{
|
||||
type: 'text',
|
||||
text,
|
||||
},
|
||||
]
|
||||
|
||||
for (const base64Image of images) {
|
||||
// Determine MIME type from base64 prefix
|
||||
let mimeType = 'image/jpeg'
|
||||
if (base64Image.includes('data:image/png')) {
|
||||
mimeType = 'image/png'
|
||||
} else if (base64Image.includes('data:image/gif')) {
|
||||
mimeType = 'image/gif'
|
||||
} else if (base64Image.includes('data:image/webp')) {
|
||||
mimeType = 'image/webp'
|
||||
}
|
||||
|
||||
// Extract base64 data (remove data URL prefix if present)
|
||||
const imageData = base64Image.includes(',')
|
||||
? base64Image.split(',')[1]
|
||||
: base64Image
|
||||
|
||||
content.push({
|
||||
type: 'image',
|
||||
image: imageData,
|
||||
mimeType,
|
||||
})
|
||||
}
|
||||
|
||||
return content
|
||||
}
|
||||
|
||||
@ -69,29 +69,41 @@ export function ChatInterface({
|
||||
const sessionKey = `chat-session-${agent.id}`
|
||||
let existingSessionId = localStorage.getItem(sessionKey)
|
||||
|
||||
// Only create new session if one doesn't exist
|
||||
if (!existingSessionId) {
|
||||
// Generate new sessionID using timestamp and random string
|
||||
existingSessionId = `session-${agent.id}-${Date.now()}-${Math.random().toString(36).substring(2, 15)}`
|
||||
localStorage.setItem(sessionKey, existingSessionId)
|
||||
}
|
||||
|
||||
setSessionId(existingSessionId)
|
||||
// Only update sessionId state if it's different from current
|
||||
setSessionId((currentSessionId) => {
|
||||
if (currentSessionId !== existingSessionId) {
|
||||
console.log(`[chat] Session changed for ${agent.id}: ${currentSessionId} -> ${existingSessionId}`)
|
||||
return existingSessionId
|
||||
}
|
||||
return currentSessionId
|
||||
})
|
||||
|
||||
// Load existing messages for this agent
|
||||
// Load existing messages for this agent only if we don't have any messages loaded
|
||||
// or if the agent ID changed
|
||||
const messagesKey = `chat-messages-${agent.id}`
|
||||
const savedMessages = localStorage.getItem(messagesKey)
|
||||
if (savedMessages) {
|
||||
try {
|
||||
const parsed = JSON.parse(savedMessages)
|
||||
// Ensure timestamps are Date objects
|
||||
const messages = parsed.map((msg: any) => ({
|
||||
const loadedMessages = parsed.map((msg: any) => ({
|
||||
...msg,
|
||||
timestamp: new Date(msg.timestamp),
|
||||
}))
|
||||
setMessages(messages)
|
||||
setMessages(loadedMessages)
|
||||
} catch (err) {
|
||||
console.error("[chat] Failed to load saved messages:", err)
|
||||
setMessages([])
|
||||
}
|
||||
} else {
|
||||
setMessages([])
|
||||
}
|
||||
}, [agent.id])
|
||||
|
||||
@ -221,39 +233,147 @@ export function ChatInterface({
|
||||
body: JSON.stringify(payload),
|
||||
})
|
||||
|
||||
const data = (await response.json()) as {
|
||||
error?: string
|
||||
hint?: string
|
||||
response?: string
|
||||
message?: string
|
||||
toolCall?: ToolCall
|
||||
if (!response.ok) {
|
||||
// Handle error response
|
||||
const errorText = await response.text()
|
||||
let errorData: { error?: string; hint?: string }
|
||||
try {
|
||||
errorData = errorText ? JSON.parse(errorText) : {}
|
||||
} catch {
|
||||
errorData = { error: errorText || "Unknown error" }
|
||||
}
|
||||
|
||||
if (!response.ok) {
|
||||
const errorMessage: Message = {
|
||||
id: (Date.now() + 1).toString(),
|
||||
role: "assistant",
|
||||
content: data.error || "Failed to communicate with the webhook.",
|
||||
content: errorData.error || "Failed to communicate with the webhook.",
|
||||
timestamp: new Date(),
|
||||
isError: true,
|
||||
hint: data.hint,
|
||||
hint: errorData.hint,
|
||||
}
|
||||
setMessages((prev) => [...prev, errorMessage])
|
||||
} else {
|
||||
// Check if this is a tool call (e.g., agent package creation)
|
||||
if (data.toolCall && data.toolCall.name === "create_agent_package") {
|
||||
const payload = data.toolCall.payload as AgentPackagePayload
|
||||
setAgentPackage(payload)
|
||||
// Don't add a regular message, the AgentForgeCard will be rendered instead
|
||||
} else {
|
||||
const assistantMessage: Message = {
|
||||
setIsLoading(false)
|
||||
return
|
||||
}
|
||||
|
||||
// Check if response has a body to stream
|
||||
if (!response.body) {
|
||||
const errorMessage: Message = {
|
||||
id: (Date.now() + 1).toString(),
|
||||
role: "assistant",
|
||||
content: data.response || data.message || JSON.stringify(data),
|
||||
content: "No response received from the webhook.",
|
||||
timestamp: new Date(),
|
||||
isError: true,
|
||||
}
|
||||
setMessages((prev) => [...prev, errorMessage])
|
||||
setIsLoading(false)
|
||||
return
|
||||
}
|
||||
|
||||
// Stream the response using plain text deltas from Vercel AI SDK
|
||||
const reader = response.body.getReader()
|
||||
const decoder = new TextDecoder()
|
||||
let accumulatedContent = ""
|
||||
let updateTimeout: NodeJS.Timeout | null = null
|
||||
const assistantMessageId = (Date.now() + 1).toString()
|
||||
|
||||
console.log("[chat] Starting to read stream...")
|
||||
|
||||
// Create initial empty assistant message
|
||||
const initialMessage: Message = {
|
||||
id: assistantMessageId,
|
||||
role: "assistant",
|
||||
content: "",
|
||||
timestamp: new Date(),
|
||||
}
|
||||
setMessages((prev) => [...prev, assistantMessage])
|
||||
setMessages((prev) => [...prev, initialMessage])
|
||||
|
||||
// Helper to batch UI updates (debounced)
|
||||
const scheduleUpdate = () => {
|
||||
if (updateTimeout) clearTimeout(updateTimeout)
|
||||
updateTimeout = setTimeout(() => {
|
||||
setMessages((prev) =>
|
||||
prev.map((msg) =>
|
||||
msg.id === assistantMessageId
|
||||
? { ...msg, content: accumulatedContent }
|
||||
: msg
|
||||
)
|
||||
)
|
||||
}, 50) // Update every 50ms instead of every chunk
|
||||
}
|
||||
|
||||
try {
|
||||
while (true) {
|
||||
const { done, value } = await reader.read()
|
||||
|
||||
if (done) {
|
||||
// Final update
|
||||
if (updateTimeout) clearTimeout(updateTimeout)
|
||||
setMessages((prev) =>
|
||||
prev.map((msg) =>
|
||||
msg.id === assistantMessageId
|
||||
? { ...msg, content: accumulatedContent }
|
||||
: msg
|
||||
)
|
||||
)
|
||||
console.log("[chat] Stream complete. Final content length:", accumulatedContent.length)
|
||||
break
|
||||
}
|
||||
|
||||
// Decode chunk - this is a text delta from Vercel AI SDK
|
||||
const chunk = decoder.decode(value, { stream: true })
|
||||
console.log("[chat] Received chunk:", chunk.substring(0, 100))
|
||||
|
||||
// Check if chunk is JSON (tool call) or plain text
|
||||
const trimmedChunk = chunk.trim()
|
||||
if (trimmedChunk.startsWith("{") && trimmedChunk.endsWith("}")) {
|
||||
// Might be a tool call
|
||||
try {
|
||||
const parsed = JSON.parse(trimmedChunk)
|
||||
console.log("[chat] Parsed JSON:", parsed)
|
||||
|
||||
// Handle tool call format from AI SDK
|
||||
if (parsed.type === "tool-call" && parsed.toolName === "create_agent_package") {
|
||||
console.log("[chat] Tool call: create_agent_package")
|
||||
const payload = parsed.toolInput
|
||||
setAgentPackage(payload)
|
||||
// Remove the accumulating message if it's empty
|
||||
if (!accumulatedContent.trim()) {
|
||||
setMessages((prev) => prev.filter((msg) => msg.id !== assistantMessageId))
|
||||
}
|
||||
}
|
||||
} catch {
|
||||
// Not valid JSON, treat as plain text
|
||||
accumulatedContent += chunk
|
||||
console.log("[chat] Text accumulation, length:", accumulatedContent.length)
|
||||
scheduleUpdate()
|
||||
}
|
||||
} else {
|
||||
// Plain text delta - accumulate it directly
|
||||
accumulatedContent += chunk
|
||||
console.log("[chat] Text accumulation, length:", accumulatedContent.length)
|
||||
scheduleUpdate()
|
||||
}
|
||||
}
|
||||
|
||||
// Stream complete - clear loading state
|
||||
setIsLoading(false)
|
||||
inputRef.current?.focus()
|
||||
} catch (streamError) {
|
||||
console.error("[chat] Stream reading error:", streamError)
|
||||
// Update message to show error
|
||||
setMessages((prev) =>
|
||||
prev.map((msg) =>
|
||||
msg.id === assistantMessageId
|
||||
? {
|
||||
...msg,
|
||||
content: msg.content || "Stream interrupted. Please try again.",
|
||||
isError: true,
|
||||
}
|
||||
: msg
|
||||
)
|
||||
)
|
||||
setIsLoading(false)
|
||||
}
|
||||
} catch (error) {
|
||||
console.error("[v0] Error sending message:", error)
|
||||
|
||||
67
src/lib/agents/definitions.ts
Normal file
67
src/lib/agents/definitions.ts
Normal file
@ -0,0 +1,67 @@
|
||||
/**
|
||||
* Agent definitions for Vercel AI SDK
|
||||
* Defines all standard agents with their prompts, tools, and configuration
|
||||
*/
|
||||
|
||||
import type { AgentDefinition } from '@/lib/types'
|
||||
// TODO: Re-enable once tool typing is fixed
|
||||
// import { createAgentPackageTool } from './tools/create-agent-package'
|
||||
import { MORGAN_SYSTEM_PROMPT } from './morgan-system-prompt'
|
||||
|
||||
/**
|
||||
* Agent 1: Repoguide - Documentation and development process assistant
|
||||
*/
|
||||
const AGENT_1_DEFINITION: AgentDefinition = {
|
||||
id: 'agent-1',
|
||||
name: 'Repoguide',
|
||||
description: 'Documenting the development process.',
|
||||
systemPrompt: `You are Repoguide, an expert documentation assistant specializing in development processes and technical documentation.
|
||||
|
||||
Your role is to:
|
||||
- Help document code, architecture, and development workflows
|
||||
- Provide clear, structured documentation
|
||||
- Answer questions about project structure and conventions
|
||||
- Suggest improvements to existing documentation
|
||||
|
||||
Respond in a clear, professional manner with proper formatting.`,
|
||||
temperature: 0.7,
|
||||
maxTokens: 4096,
|
||||
}
|
||||
|
||||
/**
|
||||
* Agent 2: Morgan - System Prompt Designer and Custom Agent Creator
|
||||
*/
|
||||
const AGENT_2_DEFINITION: AgentDefinition = {
|
||||
id: 'agent-2',
|
||||
name: 'Morgan',
|
||||
description: 'System Prompt Designer',
|
||||
systemPrompt: MORGAN_SYSTEM_PROMPT,
|
||||
// TODO: Fix tool type issue and re-enable create_agent_package tool
|
||||
// tools: {
|
||||
// create_agent_package: createAgentPackageTool,
|
||||
// },
|
||||
temperature: 0.8,
|
||||
maxTokens: 2048,
|
||||
}
|
||||
|
||||
/**
|
||||
* All standard agents indexed by ID
|
||||
*/
|
||||
export const STANDARD_AGENTS: Record<string, AgentDefinition> = {
|
||||
'agent-1': AGENT_1_DEFINITION,
|
||||
'agent-2': AGENT_2_DEFINITION,
|
||||
}
|
||||
|
||||
/**
|
||||
* Get a standard agent definition by ID
|
||||
*/
|
||||
export function getStandardAgent(agentId: string): AgentDefinition | null {
|
||||
return STANDARD_AGENTS[agentId] || null
|
||||
}
|
||||
|
||||
/**
|
||||
* Get all standard agent definitions
|
||||
*/
|
||||
export function getAllStandardAgents(): AgentDefinition[] {
|
||||
return Object.values(STANDARD_AGENTS)
|
||||
}
|
||||
88
src/lib/agents/factory.ts
Normal file
88
src/lib/agents/factory.ts
Normal file
@ -0,0 +1,88 @@
|
||||
/**
|
||||
* Agent factory for loading and managing agent definitions
|
||||
* Handles both standard agents and custom agents
|
||||
*/
|
||||
|
||||
import type { AgentDefinition } from '@/lib/types'
|
||||
import type { PinnedAgent } from '@/lib/types'
|
||||
import { getStandardAgent, STANDARD_AGENTS } from './definitions'
|
||||
|
||||
/**
|
||||
* Load an agent definition by ID
|
||||
* Supports both standard agents (agent-N) and custom agents (custom-{uuid})
|
||||
*
|
||||
* @param agentId - The agent ID to load
|
||||
* @param customAgentData - Optional custom agent data from request (for custom agents)
|
||||
* @returns The agent definition, or throws an error if not found
|
||||
*/
|
||||
export async function getAgentDefinition(
|
||||
agentId: string,
|
||||
customAgentData?: {
|
||||
systemPrompt: string
|
||||
tools?: AgentDefinition['tools']
|
||||
}
|
||||
): Promise<AgentDefinition> {
|
||||
// Check if it's a standard agent
|
||||
const standardAgent = getStandardAgent(agentId)
|
||||
if (standardAgent) {
|
||||
return standardAgent
|
||||
}
|
||||
|
||||
// Check if it's a custom agent
|
||||
if (agentId.startsWith('custom-')) {
|
||||
if (!customAgentData) {
|
||||
throw new Error(
|
||||
`Custom agent ${agentId} requires systemPrompt in request body`
|
||||
)
|
||||
}
|
||||
|
||||
// Build a custom agent definition from the provided data
|
||||
const customAgent: AgentDefinition = {
|
||||
id: agentId,
|
||||
name: agentId, // Will be overridden by client-side data
|
||||
description: 'Custom agent',
|
||||
systemPrompt: customAgentData.systemPrompt,
|
||||
tools: customAgentData.tools,
|
||||
temperature: 0.7,
|
||||
maxTokens: 4096,
|
||||
}
|
||||
|
||||
return customAgent
|
||||
}
|
||||
|
||||
throw new Error(`Agent not found: ${agentId}`)
|
||||
}
|
||||
|
||||
/**
|
||||
* Get all available standard agents for the client
|
||||
* Used by /api/agents endpoint
|
||||
*/
|
||||
export function getAllAvailableAgents(): Array<{
|
||||
id: string
|
||||
name: string
|
||||
description: string
|
||||
}> {
|
||||
return Object.values(STANDARD_AGENTS).map((agent) => ({
|
||||
id: agent.id,
|
||||
name: agent.name,
|
||||
description: agent.description,
|
||||
}))
|
||||
}
|
||||
|
||||
/**
|
||||
* Validate that an agent exists and is accessible
|
||||
*/
|
||||
export function agentExists(agentId: string): boolean {
|
||||
// Standard agents
|
||||
if (getStandardAgent(agentId)) {
|
||||
return true
|
||||
}
|
||||
|
||||
// Custom agents exist server-side if they're created by the user
|
||||
// The validation happens when systemPrompt is provided
|
||||
if (agentId.startsWith('custom-')) {
|
||||
return true // Will be validated when loading
|
||||
}
|
||||
|
||||
return false
|
||||
}
|
||||
3661
src/lib/agents/morgan-system-prompt.ts
Normal file
3661
src/lib/agents/morgan-system-prompt.ts
Normal file
File diff suppressed because it is too large
Load Diff
72
src/lib/agents/tools/create-agent-package.ts
Normal file
72
src/lib/agents/tools/create-agent-package.ts
Normal file
@ -0,0 +1,72 @@
|
||||
/**
|
||||
* Tool for Morgan agent to create custom AI agent packages
|
||||
*/
|
||||
|
||||
import { tool } from 'ai'
|
||||
import { z } from 'zod'
|
||||
import { v4 as uuidv4 } from 'uuid'
|
||||
import type { AgentPackagePayload } from '@/lib/types'
|
||||
|
||||
const createAgentPackageSchema = z.object({
|
||||
displayName: z
|
||||
.string()
|
||||
.describe('Clear, memorable name for the agent (e.g., "Code Reviewer", "API Designer")'),
|
||||
summary: z
|
||||
.string()
|
||||
.describe(
|
||||
'Concise 1-2 sentence description of what the agent does and when to use it'
|
||||
),
|
||||
systemPrompt: z
|
||||
.string()
|
||||
.describe(
|
||||
'Comprehensive system prompt that defines the agent\'s expertise, behavior, and response patterns. Should be detailed and actionable.'
|
||||
),
|
||||
tags: z
|
||||
.array(z.string())
|
||||
.describe(
|
||||
'Relevant tags categorizing the agent (e.g., ["code-review", "documentation", "analysis"])'
|
||||
),
|
||||
recommendedIcon: z
|
||||
.string()
|
||||
.optional()
|
||||
.describe('Emoji or icon identifier for visual representation'),
|
||||
whenToUse: z
|
||||
.string()
|
||||
.optional()
|
||||
.describe(
|
||||
'Guidance on when users should select this agent vs other agents'
|
||||
),
|
||||
})
|
||||
|
||||
/**
|
||||
* Morgan's create_agent_package tool
|
||||
* Enables Morgan to create custom agents with specified configurations
|
||||
*/
|
||||
export const createAgentPackageTool = tool({
|
||||
description:
|
||||
'Create a new custom AI agent with a specialized system prompt and capabilities',
|
||||
parameters: createAgentPackageSchema,
|
||||
execute: async (params: any) => {
|
||||
// Generate unique agent ID
|
||||
const agentId = `custom-${uuidv4()}`
|
||||
|
||||
// Create agent package payload
|
||||
const payload: AgentPackagePayload = {
|
||||
agentId,
|
||||
displayName: params.displayName,
|
||||
summary: params.summary,
|
||||
systemPrompt: params.systemPrompt,
|
||||
tags: params.tags || [],
|
||||
hints: {
|
||||
recommendedIcon: params.recommendedIcon,
|
||||
whenToUse: params.whenToUse,
|
||||
},
|
||||
}
|
||||
|
||||
return {
|
||||
success: true,
|
||||
message: `Created agent package "${params.displayName}" with ID: ${agentId}`,
|
||||
payload,
|
||||
}
|
||||
},
|
||||
} as any)
|
||||
107
src/lib/agents/tools/qdrant-rag.ts
Normal file
107
src/lib/agents/tools/qdrant-rag.ts
Normal file
@ -0,0 +1,107 @@
|
||||
/**
|
||||
* Qdrant RAG tool for searching knowledge base
|
||||
* Optional: Only used if QDRANT_URL and QDRANT_API_KEY are configured
|
||||
*/
|
||||
|
||||
import { tool } from 'ai'
|
||||
import { z } from 'zod'
|
||||
import { getEmbeddingModel } from '@/lib/openrouter'
|
||||
import { embed } from 'ai'
|
||||
|
||||
/**
|
||||
* Qdrant RAG tool for searching the knowledge base
|
||||
* Requires Qdrant to be configured in environment variables
|
||||
*/
|
||||
// TODO: Fix tool typing issue with Vercel AI SDK
|
||||
// Currently disabled due to strict typing in Vercel AI SDK tool() function
|
||||
export const qdrantRagTool = tool({
|
||||
description:
|
||||
'Search the knowledge base for relevant information and context. Use this to retrieve documents that can inform your responses.',
|
||||
parameters: z.object({
|
||||
query: z
|
||||
.string()
|
||||
.describe('The search query or topic to find relevant documents for'),
|
||||
topK: z
|
||||
.number()
|
||||
.int()
|
||||
.min(1)
|
||||
.max(20)
|
||||
.default(5)
|
||||
.describe('Maximum number of results to return'),
|
||||
threshold: z
|
||||
.number()
|
||||
.min(0)
|
||||
.max(1)
|
||||
.default(0.7)
|
||||
.describe('Similarity threshold (0-1) for filtering results'),
|
||||
}),
|
||||
execute: async (params: any) => {
|
||||
const { query, topK, threshold } = params
|
||||
// Check if Qdrant is configured
|
||||
const qdrantUrl = process.env.QDRANT_URL
|
||||
const qdrantApiKey = process.env.QDRANT_API_KEY
|
||||
|
||||
if (!qdrantUrl || !qdrantApiKey) {
|
||||
return {
|
||||
success: false,
|
||||
message: 'Qdrant is not configured. RAG search unavailable.',
|
||||
results: [],
|
||||
}
|
||||
}
|
||||
|
||||
try {
|
||||
// Get embedding for the query
|
||||
const { embedding } = await embed({
|
||||
model: getEmbeddingModel(),
|
||||
value: query,
|
||||
})
|
||||
|
||||
// Query Qdrant
|
||||
// Note: This is a simplified implementation
|
||||
// In production, you would use the Qdrant JS client
|
||||
const response = await fetch(`${qdrantUrl}/collections/documents/points/search`, {
|
||||
method: 'POST',
|
||||
headers: {
|
||||
'api-key': qdrantApiKey,
|
||||
'Content-Type': 'application/json',
|
||||
},
|
||||
body: JSON.stringify({
|
||||
vector: embedding,
|
||||
limit: topK,
|
||||
score_threshold: threshold,
|
||||
with_payload: true,
|
||||
}),
|
||||
})
|
||||
|
||||
if (!response.ok) {
|
||||
return {
|
||||
success: false,
|
||||
message: `Qdrant search failed: ${response.statusText}`,
|
||||
results: [],
|
||||
}
|
||||
}
|
||||
|
||||
const data = (await response.json()) as any
|
||||
|
||||
// Format results
|
||||
const results = data.result?.map((hit: any) => ({
|
||||
content: hit.payload?.text || hit.payload?.content || '',
|
||||
score: hit.score,
|
||||
source: hit.payload?.source || 'unknown',
|
||||
metadata: hit.payload,
|
||||
})) || []
|
||||
|
||||
return {
|
||||
success: true,
|
||||
message: `Found ${results.length} relevant documents`,
|
||||
results,
|
||||
}
|
||||
} catch (error) {
|
||||
return {
|
||||
success: false,
|
||||
message: `RAG search error: ${error instanceof Error ? error.message : 'Unknown error'}`,
|
||||
results: [],
|
||||
}
|
||||
}
|
||||
},
|
||||
} as any)
|
||||
49
src/lib/openrouter.ts
Normal file
49
src/lib/openrouter.ts
Normal file
@ -0,0 +1,49 @@
|
||||
/**
|
||||
* OpenRouter client configuration for Vercel AI SDK
|
||||
* Handles initialization of the OpenRouter provider and model selection
|
||||
*/
|
||||
|
||||
import { createOpenRouter } from '@openrouter/ai-sdk-provider'
|
||||
|
||||
/**
|
||||
* Get the OpenRouter client instance
|
||||
* Uses OPENROUTER_API_KEY from environment variables
|
||||
*/
|
||||
function getOpenRouterClient() {
|
||||
const apiKey = process.env.OPENROUTER_API_KEY
|
||||
if (!apiKey) {
|
||||
throw new Error('OPENROUTER_API_KEY environment variable is not set')
|
||||
}
|
||||
|
||||
return createOpenRouter({
|
||||
apiKey,
|
||||
baseURL: 'https://openrouter.ai/api/v1',
|
||||
})
|
||||
}
|
||||
|
||||
/**
|
||||
* Get the configured language model from environment
|
||||
* Falls back to gpt-oss-120b if OPENROUTER_MODEL is not set
|
||||
*/
|
||||
export function getConfiguredModel() {
|
||||
const modelId = process.env.OPENROUTER_MODEL || 'openai/gpt-oss-120b'
|
||||
const client = getOpenRouterClient()
|
||||
return client(modelId)
|
||||
}
|
||||
|
||||
/**
|
||||
* Get a specific model by ID from OpenRouter
|
||||
* Useful for overriding the default model
|
||||
*/
|
||||
export function getModelById(modelId: string) {
|
||||
const client = getOpenRouterClient()
|
||||
return client(modelId)
|
||||
}
|
||||
|
||||
/**
|
||||
* Get the embedding model for RAG (text-embedding-3-large)
|
||||
*/
|
||||
export function getEmbeddingModel() {
|
||||
const client = getOpenRouterClient()
|
||||
return client.textEmbeddingModel('openai/text-embedding-3-large')
|
||||
}
|
||||
@ -2,8 +2,25 @@
|
||||
* Core type definitions for the multi-agent chat application
|
||||
*/
|
||||
|
||||
import type { LanguageModel } from 'ai'
|
||||
import type { Tool } from 'ai'
|
||||
|
||||
/**
|
||||
* Represents an AI agent that users can chat with
|
||||
* Agent definition for Vercel AI SDK (internal - server-side)
|
||||
* Defines an agent with system prompt, tools, and LLM parameters
|
||||
*/
|
||||
export interface AgentDefinition {
|
||||
id: string
|
||||
name: string
|
||||
description: string
|
||||
systemPrompt: string
|
||||
tools?: Record<string, Tool<any, any>>
|
||||
temperature?: number
|
||||
maxTokens?: number
|
||||
}
|
||||
|
||||
/**
|
||||
* Represents an AI agent that users can chat with (client-side)
|
||||
*/
|
||||
export interface Agent {
|
||||
id: string
|
||||
|
||||
@ -14,6 +14,13 @@
|
||||
],
|
||||
"route": "agents.nicholai.work",
|
||||
"vars": {
|
||||
// LLM Configuration (Vercel AI SDK)
|
||||
"OPENROUTER_API_KEY": "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",
|
||||
"OPENROUTER_MODEL": "openai/gpt-oss-120b",
|
||||
// RAG Configuration (Qdrant)
|
||||
"QDRANT_URL": "",
|
||||
"QDRANT_API_KEY": "",
|
||||
// Agent Configuration (legacy - can be removed after migration)
|
||||
"AGENT_1_URL": "https://n8n.biohazardvfx.com/webhook/d2ab4653-a107-412c-a905-ccd80e5b76cd",
|
||||
"AGENT_1_NAME": "Repoguide",
|
||||
"AGENT_1_DESCRIPTION": "Documenting the development process.",
|
||||
@ -21,6 +28,7 @@
|
||||
"AGENT_2_NAME": "Morgan",
|
||||
"AGENT_2_DESCRIPTION": "System Prompt Designer",
|
||||
"CUSTOM_AGENT_WEBHOOK": "https://n8n.biohazardvfx.com/webhook/7cbdc539-526f-425f-abea-0886ec4c1e76",
|
||||
// Feature Flags
|
||||
"IMAGE_UPLOADS_ENABLED": "true",
|
||||
"DIFF_TOOL_ENABLED": "true"
|
||||
},
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user