Nicholai/jan
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fixed browseruse page and updated product pages

Browse Source
This commit is contained in:
Ramon Perez 2025-07-29 15:30:47 +10:00
parent a48925f6bc
commit 4b24176d2c
15 changed files with 168 additions and 1808 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

306
website/PRODUCTS_PAGE_README.MD
View File

@ -1,306 +0,0 @@
# Jan Products Page - Retro-Tech Design System
A futuristic products showcase page that combines the nostalgic aesthetics of RadioShack's electronic components with PostHog's modern data visualization approach. This creates a unique "retro-tech" visual language that feels both familiar and cutting-edge.
## Design Philosophy
### RadioShack + PostHog = Retro-Tech
- **RadioShack Heritage**: Circuit board patterns, electronic components, terminal interfaces, amber/green CRT colors
- **PostHog Influence**: Clean data cards, status indicators, system metrics, dashboard layouts
- **Result**: A cohesive aesthetic that makes AI products feel like sophisticated electronic equipment
### Color Palette
```css
--retro-green: #00ff41 /* Matrix green, primary brand */
--retro-amber: #ffb000 /* Classic terminal amber */
--retro-blue: #00bfff /* Processing/loading states */
--retro-red: #ff0040 /* Error/warning states */
--retro-cyan: #00ffff /* Accent highlights */
--retro-magenta: #ff00ff /* Special effects */
```
## Architecture
### File Structure
```
jan/website/src/
├── pages/
│ └── prods.astro # Main products page
├── components/
│ ├── CircuitBackground.astro # Animated circuit patterns
│ ├── StatusIndicator.astro # System status components
│ ├── RetroCard.astro # Terminal-style cards
│ └── FloatingNav.astro # Floating navigation menu
├── layouts/
│ └── Layout.astro # Base layout with GSAP
└── styles/
└── retro-effects.css # Advanced visual effects
```
### Component Hierarchy
```
Layout
└── Products Page
├── FloatingNav (right-side navigation)
├── Hero Section
│ └── CircuitBackground (animated)
├── Table of Contents
├── Models Section
│ └── RetroCard (circuit variant)
├── Platforms Section
│ └── RetroCard (data/system variants)
├── Tools Section
│ └── RetroCard (terminal variant)
└── Integration Section
```
## Components Deep Dive
### CircuitBackground.astro
Creates animated circuit board patterns with electronic components.
**Props:**
- `variant`: 'subtle' | 'prominent' | 'animated'
- `color`: 'green' | 'amber' | 'blue' | 'multi'
- `density`: 'low' | 'medium' | 'high'
- `animated`: boolean
- `opacity`: number (0-1)
**Features:**
- Procedural circuit patterns
- Animated electronic components (LEDs, resistors, capacitors)
- Flowing data lines
- Responsive grid scaling
### StatusIndicator.astro
Terminal-style system status indicators with pulsing animations.
**Props:**
- `status`: 'active' | 'warning' | 'success' | 'error' | 'idle'
- `label`: string
- `pulse`: boolean
- `size`: 'small' | 'medium' | 'large'
**Features:**
- Animated status dots with glow effects
- Color-coded status states
- Accessibility compliant
- Reduced motion support
### RetroCard.astro
Terminal-inspired cards with scanlines, CRT effects, and interactive animations.
**Props:**
- `variant`: 'terminal' | 'circuit' | 'data' | 'system'
- `color`: 'green' | 'amber' | 'blue' | 'red' | 'multi'
- `size`: 'small' | 'medium' | 'large'
- `glitch`: boolean
- `scanlines`: boolean
- `glow`: boolean
- `interactive`: boolean
- `status`: 'online' | 'offline' | 'error' | 'warning' | 'processing'
**Features:**
- CRT monitor bezels with power indicators
- Animated scanlines and glitch effects
- Terminal-style headers with traffic light controls
- Interactive hover states with 3D transforms
- Corner brackets for retro-futuristic framing
### FloatingNav.astro
Floating navigation with terminal interface and system monitoring.
**Props:**
- `sections`: Array of navigation items with icons and status
- `position`: 'left' | 'right'
- `theme`: 'dark' | 'matrix' | 'circuit'
- `compact`: boolean
**Features:**
- Terminal-style command prompt
- Real-time system status indicators
- Animated progress bars
- Circuit connection lines
- Smooth scrolling with GSAP integration
## Animations & Effects
### GSAP Integration
The page uses GSAP (GreenSock Animation Platform) for sophisticated animations:
**Timeline Animations:**
- Hero section entrance with staggered reveals
- Section headers with typewriter effects
- Card appearances with 3D transforms
**Scroll Triggers:**
- Elements animate into view as user scrolls
- Parallax effects on background elements
- Progressive content revelation
**Interactive Animations:**
- Magnetic cursor effects on cards
- 3D hover transformations
- Ripple effects on clicks
### CSS-Only Effects
Enhanced with pure CSS animations for performance:
**Visual Effects:**
- CRT scanlines and monitor bezels
- Matrix-style character rain
- Glitch text distortions
- Neon glow and pulsing lights
**Component Animations:**
- Electronic component blinking (LEDs)
- Data stream flowing effects
- Circuit trace pulsing
- Status indicator breathing
## Performance Optimizations
### Lazy Loading
- GSAP libraries loaded from CDN
- Animations initialized after DOM ready
- Circuit patterns use CSS transforms for GPU acceleration
### Reduced Motion Support
```css
@media (prefers-reduced-motion: reduce) {
/* All animations disabled for accessibility */
}
```
### Mobile Optimizations
- Responsive grid layouts
- Touch-friendly interactive elements
- Optimized animation performance
- Reduced visual effects on smaller screens
## Accessibility Features
### WCAG Compliance
- High contrast mode support
- Focus indicators for keyboard navigation
- Screen reader friendly markup
- Semantic HTML structure
### Keyboard Navigation
- Tab order follows logical flow
- Escape key closes floating navigation
- Arrow keys for menu navigation
- Enter/Space for activation
### Color Accessibility
- Sufficient contrast ratios maintained
- Status information not solely color-dependent
- Alternative text for decorative elements
## Customization Guide
### Adding New Sections
1. Add section to main page
2. Update FloatingNav sections array
3. Add corresponding GSAP animations
4. Update TOC links
### Creating New Card Variants
```astro
<RetroCard
variant="custom"
color="blue"
title="New Product"
subtitle="Description"
status="active"
>
<!-- Content -->
</RetroCard>
```
### Custom Circuit Patterns
```astro
<CircuitBackground
variant="prominent"
color="amber"
density="high"
animated={true}
opacity={0.2}
/>
```
### Color Theme Customization
Override CSS custom properties:
```css
:root {
--retro-green: #your-color;
--retro-amber: #your-color;
/* etc. */
}
```
## Dependencies
### Required
- Astro ^5.6.1
- GSAP (loaded from CDN)
- JetBrains Mono font
### Optional Enhancements
- Sharp for image optimization
- Additional GSAP plugins for advanced effects
## Browser Support
### Fully Supported
- Chrome 88+
- Firefox 85+
- Safari 14+
- Edge 88+
### Graceful Degradation
- Older browsers receive static layout
- CSS fallbacks for unsupported features
- Progressive enhancement approach
## Development Commands
```bash
# Start development server
npm run dev
# Build for production
npm run build
# Preview production build
npm run preview
```
## Future Enhancements
### Planned Features
- Sound effects for interactions
- More electronic component animations
- Additional card variants
- Theme switching capability
- Advanced glitch effects
### Performance Improvements
- WebGL circuit backgrounds
- Intersection Observer for animations
- Service worker for offline support
- Image lazy loading optimization
## Contributing
When adding new components or effects:
1. Follow the retro-tech design language
2. Ensure accessibility compliance
3. Add reduced motion alternatives
4. Test on multiple devices
5. Document new props and features
---
*Created with ❤️ and a healthy dose of 80s nostalgia*

306
website/PRODUCTS_PAGE_README.md
View File

@ -1,306 +0,0 @@
# Jan Products Page - Retro-Tech Design System
A futuristic products showcase page that combines the nostalgic aesthetics of RadioShack's electronic components with PostHog's modern data visualization approach. This creates a unique "retro-tech" visual language that feels both familiar and cutting-edge.
## Design Philosophy
### RadioShack + PostHog = Retro-Tech
- **RadioShack Heritage**: Circuit board patterns, electronic components, terminal interfaces, amber/green CRT colors
- **PostHog Influence**: Clean data cards, status indicators, system metrics, dashboard layouts
- **Result**: A cohesive aesthetic that makes AI products feel like sophisticated electronic equipment
### Color Palette
```css
--retro-green: #00ff41 /* Matrix green, primary brand */
--retro-amber: #ffb000 /* Classic terminal amber */
--retro-blue: #00bfff /* Processing/loading states */
--retro-red: #ff0040 /* Error/warning states */
--retro-cyan: #00ffff /* Accent highlights */
--retro-magenta: #ff00ff /* Special effects */
```
## Architecture
### File Structure
```
jan/website/src/
├── pages/
│ └── prods.astro # Main products page
├── components/
│ ├── CircuitBackground.astro # Animated circuit patterns
│ ├── StatusIndicator.astro # System status components
│ ├── RetroCard.astro # Terminal-style cards
│ └── FloatingNav.astro # Floating navigation menu
├── layouts/
│ └── Layout.astro # Base layout with GSAP
└── styles/
└── retro-effects.css # Advanced visual effects
```
### Component Hierarchy
```
Layout
└── Products Page
├── FloatingNav (right-side navigation)
├── Hero Section
│ └── CircuitBackground (animated)
├── Table of Contents
├── Models Section
│ └── RetroCard (circuit variant)
├── Platforms Section
│ └── RetroCard (data/system variants)
├── Tools Section
│ └── RetroCard (terminal variant)
└── Integration Section
```
## Components Deep Dive
### CircuitBackground.astro
Creates animated circuit board patterns with electronic components.
**Props:**
- `variant`: 'subtle' | 'prominent' | 'animated'
- `color`: 'green' | 'amber' | 'blue' | 'multi'
- `density`: 'low' | 'medium' | 'high'
- `animated`: boolean
- `opacity`: number (0-1)
**Features:**
- Procedural circuit patterns
- Animated electronic components (LEDs, resistors, capacitors)
- Flowing data lines
- Responsive grid scaling
### StatusIndicator.astro
Terminal-style system status indicators with pulsing animations.
**Props:**
- `status`: 'active' | 'warning' | 'success' | 'error' | 'idle'
- `label`: string
- `pulse`: boolean
- `size`: 'small' | 'medium' | 'large'
**Features:**
- Animated status dots with glow effects
- Color-coded status states
- Accessibility compliant
- Reduced motion support
### RetroCard.astro
Terminal-inspired cards with scanlines, CRT effects, and interactive animations.
**Props:**
- `variant`: 'terminal' | 'circuit' | 'data' | 'system'
- `color`: 'green' | 'amber' | 'blue' | 'red' | 'multi'
- `size`: 'small' | 'medium' | 'large'
- `glitch`: boolean
- `scanlines`: boolean
- `glow`: boolean
- `interactive`: boolean
- `status`: 'online' | 'offline' | 'error' | 'warning' | 'processing'
**Features:**
- CRT monitor bezels with power indicators
- Animated scanlines and glitch effects
- Terminal-style headers with traffic light controls
- Interactive hover states with 3D transforms
- Corner brackets for retro-futuristic framing
### FloatingNav.astro
Floating navigation with terminal interface and system monitoring.
**Props:**
- `sections`: Array of navigation items with icons and status
- `position`: 'left' | 'right'
- `theme`: 'dark' | 'matrix' | 'circuit'
- `compact`: boolean
**Features:**
- Terminal-style command prompt
- Real-time system status indicators
- Animated progress bars
- Circuit connection lines
- Smooth scrolling with GSAP integration
## Animations & Effects
### GSAP Integration
The page uses GSAP (GreenSock Animation Platform) for sophisticated animations:
**Timeline Animations:**
- Hero section entrance with staggered reveals
- Section headers with typewriter effects
- Card appearances with 3D transforms
**Scroll Triggers:**
- Elements animate into view as user scrolls
- Parallax effects on background elements
- Progressive content revelation
**Interactive Animations:**
- Magnetic cursor effects on cards
- 3D hover transformations
- Ripple effects on clicks
### CSS-Only Effects
Enhanced with pure CSS animations for performance:
**Visual Effects:**
- CRT scanlines and monitor bezels
- Matrix-style character rain
- Glitch text distortions
- Neon glow and pulsing lights
**Component Animations:**
- Electronic component blinking (LEDs)
- Data stream flowing effects
- Circuit trace pulsing
- Status indicator breathing
## Performance Optimizations
### Lazy Loading
- GSAP libraries loaded from CDN
- Animations initialized after DOM ready
- Circuit patterns use CSS transforms for GPU acceleration
### Reduced Motion Support
```css
@media (prefers-reduced-motion: reduce) {
/* All animations disabled for accessibility */
}
```
### Mobile Optimizations
- Responsive grid layouts
- Touch-friendly interactive elements
- Optimized animation performance
- Reduced visual effects on smaller screens
## Accessibility Features
### WCAG Compliance
- High contrast mode support
- Focus indicators for keyboard navigation
- Screen reader friendly markup
- Semantic HTML structure
### Keyboard Navigation
- Tab order follows logical flow
- Escape key closes floating navigation
- Arrow keys for menu navigation
- Enter/Space for activation
### Color Accessibility
- Sufficient contrast ratios maintained
- Status information not solely color-dependent
- Alternative text for decorative elements
## Customization Guide
### Adding New Sections
1. Add section to main page
2. Update FloatingNav sections array
3. Add corresponding GSAP animations
4. Update TOC links
### Creating New Card Variants
```astro
<RetroCard
variant="custom"
color="blue"
title="New Product"
subtitle="Description"
status="active"
>
<!-- Content -->
</RetroCard>
```
### Custom Circuit Patterns
```astro
<CircuitBackground
variant="prominent"
color="amber"
density="high"
animated={true}
opacity={0.2}
/>
```
### Color Theme Customization
Override CSS custom properties:
```css
:root {
--retro-green: #your-color;
--retro-amber: #your-color;
/* etc. */
}
```
## Dependencies
### Required
- Astro ^5.6.1
- GSAP (loaded from CDN)
- JetBrains Mono font
### Optional Enhancements
- Sharp for image optimization
- Additional GSAP plugins for advanced effects
## Browser Support
### Fully Supported
- Chrome 88+
- Firefox 85+
- Safari 14+
- Edge 88+
### Graceful Degradation
- Older browsers receive static layout
- CSS fallbacks for unsupported features
- Progressive enhancement approach
## Development Commands
```bash
# Start development server
npm run dev
# Build for production
npm run build
# Preview production build
npm run preview
```
## Future Enhancements
### Planned Features
- Sound effects for interactions
- More electronic component animations
- Additional card variants
- Theme switching capability
- Advanced glitch effects
### Performance Improvements
- WebGL circuit backgrounds
- Intersection Observer for animations
- Service worker for offline support
- Image lazy loading optimization
## Contributing
When adding new components or effects:
1. Follow the retro-tech design language
2. Ensure accessibility compliance
3. Add reduced motion alternatives
4. Test on multiple devices
5. Document new props and features
---
*Created with ❤️ and a healthy dose of 80s nostalgia*

BIN
website/src/assets/jan-server.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 357 KiB

162
website/src/content/products/index.mdx
View File

@ -1,13 +1,22 @@
---
title: Products
description: AI that runs where you need it, how you need it
title: Vision 🚀
description: AI that runs where you need it, how you need it.
sidebar:
order: 0
---
import { Aside, Card, CardGrid } from '@astrojs/starlight/components';
Jan is moving from a local AI application to a complete full-stack AI solution that you can self-host. This includes models, applications, and tools that delights users and help them solve their problems.
**Jan's Goal is**
> to build a superintelligence that you can self-host and use locally on your own devices.
**We know it's hard**
> but we believe this will be possible in the next decade.
Jan is moving from a local AI application to a complete full-stack AI solution that you can self-host. This includes models, applications, and tools that delight users and help them solve their problems.
## What We're Building
@ -16,26 +25,27 @@ Jan is moving from a local AI application to a complete full-stack AI solution t
Unlike other AI assistants that do specific tasks with one model or have many models with a myriad of solutions, Jan provides:
- Its own specialised models that are optimised at specific tasks like web-search, creative writing, and translation
- Applications that work across all of your devices in an integrated way
- Tools that actually get things done
- Tools to help you get things done
## Two Modes, One Experience
### Local (Incognito) Mode
## Two Modes
### Local Mode
Run AI models entirely on your device, giving you complete privacy with no internet required.
### Cloud Mode
![Jan Desktop](../../assets/jan_desktop.png)
### Cloud Mode
Connect to more powerful models when needed - either self-hosted or via jan.ai.
<Aside type="tip" title="Simple Choice">
![Jan Everywhere](../../assets/jan_everywhere.png)
<Aside type="note">
Users shouldn't need to understand models, APIs, or technical details. Just choose Local for privacy or Cloud for power.
</Aside>
## Our Product Principles
### 1) It Just Works
1. Open Jan, start chatting
2. Onboarding is fully available but optional
3. Setting up an API key is optional
@ -45,12 +55,10 @@ Users shouldn't need to understand models, APIs, or technical details. Just choo
We handle the complexity.
### 2) Cloud When Needed
Start completely locally and own your AI models. Add cloud capabilities only when you choose to.
### 3) Solve Problems, Not Settings
We help users get to answers quickly answers, not configuration options. Power users can dig deeper, but it's never required.
We help users get to answers quickly and leave the configurations as optional. Power users can dig deeper, but it's never required.
## Available on Every Device
@ -58,118 +66,41 @@ We help users get to answers quickly answers, not configuration options. Power u
<Card title="Jan Desktop" icon="laptop">
**Available Now**
Your personal AI workstation that helps with our use cases and powers other devices. Run models locally right away or bring an API key to connect to your favorite cloud-based models.
Your personal AI workstation that helps with your use cases and powers other devices. Run models locally right away or bring an API key to connect to your favorite cloud-based models.
- Runs models locally on your hardware
- GPU acceleration support
- Powers other devices via network connection
- Complete privacy and control
- Windows, macOS, and Linux support
**Requirements:**
- Minimum 8GB RAM
- 10GB+ storage space
- Optional: NVIDIA GPU for acceleration
**[Learn more &rarr;](./platforms/desktop)**
</Card>
<Card title="Jan Web" icon="up-arrow">
**Beta Launch Soon**
Web-based version of Jan with no setup required. Same default cloud mode for mobile and desktop users.
Web-based version of Jan with no setup required. The same default cloud mode for mobile and desktop users.
- No installation needed
- Instant access from any browser
- Automatic updates and maintenance
- Default cloud backend for mobile apps
- Team collaboration features
**Pricing:**
- Free for everyone
- Pro: Access our latest models
- Enterprise: Self-host or we host it for you
**[Learn more &rarr;](./platforms/jan-ai)**
</Card>
<Card title="Jan Mobile" icon="phone">
**Coming Q4 2025**
Connect to Desktop/Server, run local mode with Jan Nano or Lucy, same experience everywhere.
Connect to your Desktop/Server, or run local models like Jan Nano. The same experience, everywhere.
Jan Mobile adapts to your situation:
- **At Home**: Connect to your Jan Desktop over WiFi
- **At Work**: Connect to your company Jan Server
- **On the Go**: Run Jan Nano on your phone or talk to cloud models
- iOS and Android support
- Three adaptive modes (Desktop, Server, Local)
- Voice-first interface
- Seamless device switching
- Jan Nano for on-device AI
**[Learn more &rarr;](./platforms/mobile)**
</Card>
<Card title="Jan Server" icon="bars">
**Coming Q2 2025**
**Coming Q3 2025**
Self-hosted solution for teams and enterprises. Your own private AI cloud.
Self-hosted solution or connect to Jan via API for teams and enterprises.
- Support for 5-500+ concurrent users
- Enterprise authentication (SSO, LDAP)
- Docker and Kubernetes deployment
- Admin dashboard
- Team knowledge sharing
**Deployment Options:**
- Docker: Single command setup
- Kubernetes: Enterprise scale
- Bare metal: Maximum control
**[Learn more &rarr;](./platforms/server)**
</Card>
</CardGrid>
## Jan Mobile: Three Modes, One Experience
Jan Mobile brings the same AI experience to your phone. Connect to your desktop, your server, or run models locally.
### How It Works
Jan Mobile adapts to your situation:
**At Home** - Connect to your Jan Desktop over WiFi
Your Phone → WiFi → Your Desktop → Response
**At Work** - Connect to your company Jan Server
Your Phone → Internet → Company Server → Response
**On the Go** - Run Jan Nano on your phone or talk to your favorite cloud-based model
Your Phone → Jan Nano (6GB) → Response
No configuration needed. It just works.
### Key Features
- **Seamless Switching**: Move from home to office to airplane. One-click and Jan adapts immediately.
- **Voice First**: Talk to Jan naturally. Responses can be spoken too.
- **Sync Everything**: Conversations, settings, and preferences follow you across devices.
### Privacy & Security
**Your Data, Your Control**
- Local Mode: Everything stays on your phone
- Desktop Mode: Direct encrypted connection
- Server Mode: Your organization's policies apply
**No Compromises**
- Biometric app lock
- Encrypted storage
- No cloud backups without permission
- Clear data anytime
## What Makes Jan Different
| Feature | Other AI Assistants | Jan |
|---------|---------------------|-----|
| Models | Wrapper around Claude/GPT | Our own models + You can own them |
| Dual mode | Your data on their servers | Your data stays yours |
| Deployment | Cloud only | Local, self-hosted, or cloud |
| Cost | Subscription forever | Free locally, pay for cloud |
| :--- | :--- | :--- |
| **Models** | Wrapper around Claude/GPT | Our own models + You can own them |
| **Dual mode** | Your data on their servers | Your data stays yours |
| **Deployment** | Cloud only | Local, self-hosted, or cloud |
| **Cost** | Subscription forever | Free locally, pay for cloud |
## Development Timeline
@ -178,7 +109,7 @@ Jan is actively developed with regular releases. Our development follows these k
### Current Focus
- **Jan Desktop**: Continuous improvements and model support
- **Jan Web**: Beta launch preparation
- **Model Development**: Jan Nano and Lucy optimization
- **Model Development**: Jan Nano optimization and v1 launch
### Next 6 Months
- Jan Web public beta
@ -186,25 +117,10 @@ Jan is actively developed with regular releases. Our development follows these k
- Server deployment tools
### Future Vision
- Complete AI Agent platform
- Complete full-stack AI solution
- Advanced tool integration
- Enterprise features
<Aside type="tip">
We're building AI that respects your choices. Run it locally for privacy, connect to cloud for power, or self-host for both.
We're building AI that respects your choices. Run it locally and power other apps, connect to the cloud for power, or self-host for both.
</Aside>
---
**Ready to Get Started?**
- [Download Jan Desktop](/download) - Available now for Windows, macOS, and Linux
- [Join Jan Web Beta](/beta) - Get early access to our web platform
- [Read the Documentation](/docs) - Learn how to use Jan effectively
- [Explore Our Models](/docs/models) - Discover Jan's specialized AI models
For detailed information about each platform, explore the individual product pages:
- [Jan Desktop](/docs/products/platforms/desktop)
- [Jan Web](/docs/products/platforms/jan-ai)
- [Jan Mobile](/docs/products/platforms/mobile)
- [Jan Server](/docs/products/platforms/server)

198
website/src/content/products/models/jan-nano.mdx
View File

@ -1,207 +1,33 @@
---
title: Jan Nano
description: Compact research model optimized for finding answers
description: Compact research model optimized for finding answers through tool use.
sidebar:
order: 2
---
import { Aside, Card, CardGrid } from '@astrojs/starlight/components';
import { Aside } from '@astrojs/starlight/components';
Jan Nano is a 4-billion parameter model designed for research and information retrieval. Instead of trying to know everything, it excels at finding anything through deep integration with Model Context Protocol (MCP) tools.
## Two Variants
| Model | Context Window | Size | Use Case |
|-------|----------------|------|----------|
|:---|:---|:---|:---|
| Jan Nano 32k | 32,768 tokens | 4-8GB | Quick research, general queries |
| Jan Nano 128k | 131,072 tokens | 8-12GB | Deep research, document analysis |
<Aside type="note">
Jan Nano requires MCP-enabled tools (like web search) to reach its full potential. Enable MCP in Settings → Advanced Settings.
Jan Nano requires MCP-enabled tools (like web search) to reach its full potential. Enable MCP in **Settings → Advanced Settings**.
</Aside>
## What Makes Nano Different
### Research-First Design
Jan Nano isn't trained to memorize facts. It's trained to:
- Find relevant information quickly
- Synthesize findings from multiple sources
- Maintain context across long research sessions
- Provide accurate citations
### MCP Integration
Works seamlessly with:
- Web search (Serper, Exa)
- Document analysis
- Code repositories
- Custom data sources
### Extended Context
The 128k variant can process:
- 50+ research papers simultaneously
- Entire codebases
- Book-length documents
- Thousand-page contracts
## Performance
### Hardware Requirements
<CardGrid>
<Card title="Minimum (32k)" icon="laptop">
- 8GB RAM
- Any modern CPU
- Works on most devices
</Card>
<Card title="Recommended (128k)" icon="server">
- 16GB RAM
- GPU with 8GB+ VRAM
- CUDA-compatible
</Card>
</CardGrid>
### Speed Benchmarks
| Device | Variant | Quantization | Speed |
|--------|---------|--------------|-------|
| M2 MacBook | 32k | Q4 | 80 tokens/s |
| RTX 4090 | 32k | Q8 | 200+ tokens/s |
| M2 MacBook | 128k | Q4 | 40 tokens/s |
| RTX 4090 | 128k | Q8 | 100+ tokens/s |
## Getting Started
### 1. Enable MCP
Go to **Settings** → **Model Providers** → **Llama.cpp** and enable tool use for Jan Nano.
### 2. Add Search Tool
Add a search MCP like Serper:
- Get free API key from [serper.dev](https://serper.dev)
- Add to **Settings** → **MCP Servers**
### 3. Start Researching
Open a chat and ask Jan Nano to search for information:
```
"What are the latest developments in quantum computing?"
"Search for recent breakthroughs in mRNA vaccine technology"
"Find and compare different approaches to carbon capture"
```
## Example Queries
### Current Events
- Latest renewable energy developments in Europe
- Recent AI announcements from major tech companies
- Current economic indicators across G7 nations
### Deep Research
- Analyze 20 papers on climate change impacts
- Compare different programming paradigms with examples
- Synthesize findings from multiple medical studies
### Business Intelligence
- Market analysis across competitors
- Technology trends in specific sectors
- Regulatory changes affecting industries
<Aside type="tip">
Jan Nano shines when you need to find and synthesize information, not when you need creative writing or complex reasoning without external data.
</Aside>
- **Research-First Design:** Trained to find relevant information, synthesize findings, and provide accurate citations.
- **MCP Integration:** Works seamlessly with tools like web search, document analysis, and code repositories.
- **Extended Context:** The 128k variant can process entire codebases, book-length documents, or 50+ research papers simultaneously.
## Technical Details
- **Base:** 4B parameter transformer
- **Training:** Optimized for tool use and retrieval
- **Quantization:** Q4, Q8, FP16 variants available
### Architecture
- **Base**: 4B parameter transformer
- **Training**: Optimized for tool use and retrieval
- **Context**: Native support (not retrofitted)
- **Quantization**: Q4, Q8, FP16 variants
### Why Native Context Matters
Unlike models extended with YaRN or PI methods, Jan Nano 128k:
- Maintains consistent performance across full context
- Actually improves with more context (inverse scaling)
- No performance cliff at higher token counts
### Deployment Options
#### VLLM (Recommended for 128k)
```bash
vllm serve Menlo/Jan-nano-128k \
--host 0.0.0.0 \
--port 1234 \
--enable-auto-tool-choice \
--tool-call-parser hermes \
--rope-scaling '{"rope_type":"yarn","factor":3.2,"original_max_position_embeddings":40960}' \
--max-model-len 131072
```
#### llama.cpp
```bash
llama-server ... --rope-scaling yarn --rope-scale 3.2 --yarn-orig-ctx 40960
```
## Limitations
### Not Designed For
- Creative writing
- Complex reasoning without data
- Mathematical proofs
- Code generation from scratch
### Best Used For
- Information retrieval
- Research synthesis
- Document analysis
- Fact-finding missions
## Choosing Between Variants
### Use 32k When
- Running on limited hardware
- Need quick responses
- Researching focused topics
- Battery life matters
### Use 128k When
- Analyzing multiple documents
- Deep research projects
- Processing entire codebases
- Hardware isn't a constraint
## Integration Examples
### With Web Search
```python
# Jan Nano automatically uses search when needed
response = jan.chat(
model="jan-nano-32k",
message="Find the latest SpaceX launch details",
tools=["search"]
)
```
### With Document Analysis
```python
# Process long documents efficiently
response = jan.chat(
model="jan-nano-128k",
message="Summarize key findings from these papers",
attachments=["paper1.pdf", "paper2.pdf", "paper3.pdf"]
)
```
## The Philosophy
Most models try to be encyclopedias. Jan Nano is a research assistant. It doesn't memorize the internet - it knows how to navigate it.
This focused approach means:
- Smaller model size
- Better accuracy with sources
- More reliable information
- Efficient hardware usage
---
[Download Jan Desktop](https://jan.ai/download) | [Model Details](https://huggingface.co/Menlo/Jan-nano) | [MCP Documentation](https://jan.ai/docs/mcp)
## Philosophy
Most models try to be encyclopedias. Jan Nano is a research assistant. It doesn't memorize the internet—it knows how to navigate it.

27
website/src/content/products/models/jan-v1.mdx
View File

@ -6,39 +6,24 @@ sidebar:
banner:
content: 'In Development: Jan V1 models are currently being trained and are not yet available.'
---
import { Aside, Card, Tabs, TabItem } from '@astrojs/starlight/components';
import { Aside } from '@astrojs/starlight/components';
## Our Foundational Model Family
Jan V1 is our in-house family of models designed to compete directly with leading models like GPT-4 and Claude. We're not just fine-tuning; we're building powerful, general-purpose models from the ground up to solve real-world problems with a focus on efficiency and privacy.
<Aside type="note">
**Jan V1 models are currently in the pre-training phase.** They are not yet available for download or use. The information on this page reflects our development goals and roadmap.
</Aside>
### The Vision
Most AI applications today are simply wrappers around a few proprietary models. We believe in providing a powerful, open alternative that you can run on your own terms. Jan V1 is our commitment to that vision.
Jan V1 is our in-house, still in training, family of models designed to compete directly with leading models. We're building powerful, general-purpose models from the ground up to solve real-world problems with a focus on efficiency and privacy.
### Planned Model Lineup
| Model | Target Size | Intended Use Case | Availability |
|-------------|-------------|------------------------------|---------------|
|:------------|:------------|:-----------------------------|:--------------|
| Jan V1-7B | 4-8GB | Fast, efficient daily tasks | Coming Soon |
| Jan V1-13B | 8-16GB | Balanced power and performance | Coming Soon |
| Jan V1-70B | 40-64GB | Deep analysis, professional work | Coming Soon |
| Jan V1-180B | 100GB+ | Frontier research, complex tasks | Planned 2026 |
| Jan V1-2350B | 100GB+ | Frontier research, complex tasks | Planned 2026 |
### What to Expect
- **Competitive Performance**: Aiming for results on par with leading closed-source models.
- **Optimized for Local Use**: Efficient quantized versions for running on your own hardware.
- **Privacy-Centric**: Trainable and runnable in your own environment, ensuring your data stays yours.
- **Seamless Integration**: Designed to work perfectly within the Jan ecosystem, from Desktop to Server.
---
### Stay Updated
Be the first to know about our training progress, benchmark results, and when the first Jan V1 models are released.
[**Follow our development &rarr;**](https://jan.ai/v1-updates)
- **Seamless Integration**: Designed to work perfectly within the Jan ecosystem.
- **Fine-tuning support**: Easy to adapt to specific tasks or domains.

170
website/src/content/products/platforms/desktop.mdx
View File

@ -1,17 +1,18 @@
---
title: Jan Desktop
description: AI that runs on your computer, not someone else's
description: AI that runs on your computer, not someone else's. Your personal AI workstation.
sidebar:
order: 2
---
import { Aside, Card, CardGrid, Tabs, TabItem } from '@astrojs/starlight/components';
This is how Jan started and it has been available since day 1. Jan Desktop strives to be:
This is how Jan started and it has been available since day 1.
> Your personal AI workstation that helps with our use cases and powers other devices. Run models locally right away or bring an API key to connect to your favorite cloud-based models.
Jan Desktop strives to be:
Jan Desktop is where it all starts. Download it, open it, start chatting. Your AI runs on your computer with zero setup required.
> Your personal AI workstation that helps with your use cases and powers other devices. Run models locally right away or bring an API key to connect to your favorite cloud-based models.
Jan Desktop is where it all starts. Download it, open it, and start chatting. Your AI runs on your computer with zero setup required.
## Two Modes, Zero Complexity
@ -22,7 +23,7 @@ Your conversations stay on your computer. No internet needed. Complete privacy.
Connect to more powerful models when you need them. Your choice of provider.
<Aside type="note">
First time opening Jan? It just works. No API keys, no model downloads, no settings required.
As of today, when you first open Jan you do have to download a model or connect to a cloud provider, but that is about to change soon.
</Aside>
## What You Get
@ -31,23 +32,18 @@ First time opening Jan? It just works. No API keys, no model downloads, no setti
<Card title="Works Offline" icon="wifi">
Download once, use forever. Internet is optional.
</Card>
<Card title="Your Data Stays Yours" icon="shield">
Everything stored in `~/jan`. No cloud backups unless you want them.
Everything is stored in `~/.local/share/jan`. No cloud backups unless you want them.
</Card>
<Card title="Powers Other Devices" icon="devices">
Your desktop becomes an AI server for your phone and other computers via network connection.
Your desktop becomes an AI server for your phone and other computers.
</Card>
<Card title="Developer Friendly" icon="code">
Local API at `localhost:1337`. Works with any OpenAI-compatible tool.
</Card>
<Card title="GPU Acceleration" icon="rocket">
Automatically detects and uses NVIDIA GPUs for faster performance.
</Card>
<Card title="Cross-Platform" icon="laptop">
Windows, macOS, and Linux support with native performance.
</Card>
@ -66,51 +62,29 @@ First time opening Jan? It just works. No API keys, no model downloads, no setti
- **GPU:** NVIDIA GPU with 6GB+ VRAM for acceleration
- **OS:** Latest versions for best performance
### GPU Acceleration (Optional)
- **NVIDIA Driver:** 470.63.01 or higher
- **CUDA Toolkit:** 11.7 or higher
- **Linux:** Additional GCC-11 packages required
<Aside type="tip">
Jan automatically detects your hardware and recommends the best model. No manual configuration needed.
</Aside>
## Getting Started
1. **Download Jan** from [jan.ai/download](https://jan.ai/download)
2. **Open the app** - it loads with everything ready
3. **Start chatting** - that's it
1. **Download Jan** from [jan.ai/download](https://jan.ai/download)
2. **Open the app** - it loads with everything ready
3. **Start chatting** - that's it
## Local Mode Features
### Built-in Model
Jan comes with a model that works immediately. No downloading, no waiting.
### Smart Defaults
- Automatically uses your GPU if available
- Adjusts to your system's capabilities
- Works on older hardware too
### Complete Privacy
- No telemetry by default
- No account required
- No data leaves your machine
- Chat history stored locally only
- Full control over your data
- **Select your favorite Model:** Jan allows you to download any GGUF model from the Hugging Face Hub.
- **Smart Defaults:** Automatically uses your GPU if available and adjusts to your system's capabilities.
- **Complete Privacy:** No telemetry by default, no account required, and no data leaves your machine.
## Cloud Mode (Optional)
Connect to external AI providers when you need more power:
<Tabs syncKey="cloud-provider">
<Tabs>
<TabItem label="jan.ai">
Our cloud service (coming soon). One click to enable.
</TabItem>
<TabItem label="OpenAI">
Use your OpenAI API key for GPT-4 access.
</TabItem>
<TabItem label="Self-Hosted">
Connect to your own Jan Server.
</TabItem>
@ -118,19 +92,11 @@ Connect to external AI providers when you need more power:
## Desktop as Your AI Hub
Your desktop can power AI across all your devices. Jan Desktop automatically becomes a local server that other devices can connect to:
Your desktop can power AI across all your devices by automatically becoming a local server.
**Network Sharing:**
- Mobile apps connect over WiFi
- Other computers can access your models
- API available at `localhost:1337`
- No internet required for local network
**Supported Connections:**
- Jan Mobile (coming Q4 2025)
- Other Jan Desktop instances
- Any OpenAI-compatible application
- Custom scripts and tools
- **Network Sharing:** Mobile apps connect over WiFi, and other computers can access your models.
- **API:** Available at `localhost:1337` for any OpenAI-compatible application.
- **Offline Access:** No internet required for local network connections.
## For Developers
@ -139,101 +105,9 @@ Your desktop can power AI across all your devices. Jan Desktop automatically bec
# Always running at localhost:1337
curl http://localhost:1337/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{"model": "default", "messages": [{"role": "user", "content": "Hello"}]}'
-d '{"model": "gemma3:3b", "messages": [{"role": "user", "content": "Hello"}]}'
```
### Works With
- Continue (AI coding assistant)
- Cursor (AI editor)
- Any OpenAI SDK
- Your custom scripts
## Common Questions
### Do I need to download models?
No. Jan comes with a default model that works immediately.
### Can I use it offline?
Yes. Local Mode works completely offline once installed.
### How do I switch models?
Most users don't need to. Power users can explore the Model Hub.
### Is it really private?
Yes. In Local Mode, nothing leaves your computer.
## Performance Expectations
| Task | Local Mode | Cloud Mode |
|------|------------|------------|
| Simple questions | Instant | Instant |
| Complex analysis | 2-5 seconds | 1-2 seconds |
| Code generation | Good | Excellent |
| Privacy | Complete | Depends on provider |
## Performance Optimization
### Automatic GPU Detection
Jan automatically detects and configures:
- NVIDIA graphics cards
- Available VRAM
- Optimal GPU layers (ngl) settings
- Fallback to CPU if needed
### Memory Management
- Smart model loading based on available RAM
- Automatic offloading when memory is low
- Background process cleanup
- Efficient resource usage
## Tested Configurations
Jan Desktop has been verified to work on these systems:
**Windows Systems:**
- Windows 11 Pro + RTX 4070Ti + CUDA 12.2 + Driver 531.18
- Windows 10 + GTX 1660Ti + Various driver versions
**Linux Systems:**
- Ubuntu 22.04 + RTX 4070Ti + CUDA 12.2 + Driver 545
- Ubuntu 20.04 on Proxmox VM + GTX 1660Ti + CUDA 12.1
**macOS Systems:**
- macOS Monterey and newer
- Both Intel and Apple Silicon (M1/M2) processors
- Automatic Metal acceleration on Apple Silicon
## Coming Soon
### Enhanced User Experience
- One-click model switching
- Improved onboarding flow
- Better hardware detection
- Voice input support
### Advanced Features
- Model fine-tuning interface
- Custom model deployment
- Team collaboration features
- Enhanced mobile connectivity
<Aside type="caution">
Having issues? Most problems are solved by restarting the app. If that doesn't work, check our [troubleshooting guide](/docs/troubleshooting).
</Aside>
## Why Desktop?
Your computer is powerful. It has GPUs, CPUs, and memory that can run AI locally. Jan Desktop uses that power to give you:
- **Privacy**: Your data never leaves your machine
- **Control**: Update when you want, use what you want
- **Reliability**: Works offline, no API limits
- **Speed**: No network latency for local models
## The Bottom Line
Jan Desktop is AI that respects your computer is YOUR computer. Not a terminal to someone else's server. Just software that works for you.
---
[Download Jan Desktop](https://jan.ai/download) | [Documentation](/docs) | [Discord Community](https://discord.gg/jan)
Jan Desktop is AI that respects that your computer is YOUR computer, not a terminal to someone else's server. Just software that works for you.

167
website/src/content/products/platforms/jan-ai.mdx
View File

@ -1,189 +1,46 @@
---
title: jan.ai
description: Cloud AI that respects your privacy
title: Jan.ai
description: Cloud AI that respects your privacy. Web-based access to Jan with no setup required.
sidebar:
order: 1
---
import { Aside, Card, CardGrid } from '@astrojs/starlight/components';
**Status:** Beta Launch Soon
![Jan Web](../../../assets/jan_web.png)
Web-based version of Jan with no setup required. Same default cloud mode for mobile and desktop users.
**Status:** Beta Launch Soon 🚀
Web-based version of Jan with no setup required. Same default cloud mode for (soon) mobile and desktop users.
## What is Jan Web?
Jan Web is the cloud-hosted version of Jan that runs in your browser. No installation needed, instant access from any device, with the same AI experience you get locally.
<Aside type="note">
Currently in beta development. Beta launch coming soon.
Currently in development with a beta launch coming soon.
</Aside>
## How It Works
<CardGrid>
<Card title="For Desktop Users" icon="laptop">
When you switch to Cloud Mode in Jan Desktop, it connects to Jan Web automatically. No configuration needed.
You can sync conversations from Jan desktop to Jan Web.
</Card>
<Card title="For Mobile Users" icon="mobile">
Jan Web serves as the default cloud backend for mobile apps when not connected to your desktop.
Jan Web uses the same models you have access to on the go.
</Card>
<Card title="For Web Users" icon="globe">
Visit jan.ai directly in your browser for instant access to AI without downloading anything.
Visit (soon) ask.jan.ai directly in your browser for instant access to AI without downloading anything.
</Card>
<Card title="Team Collaboration" icon="group">
Share prompts, workflows, and collaborate on threads with your team members.
</Card>
</CardGrid>
## Key Features
### No Installation Needed
- No installation needed
- Instant access from any browser
- Automatic updates and maintenance
- Works on any device with internet
### Team Collaboration Features
- Share prompts across team members
- Share workflows and templates
- Collaborate on conversation threads
- Real-time team coordination
### Cloud Backend for Mobile
- Default cloud backend for mobile apps
- Seamless sync across devices
- Consistent experience everywhere
- No setup required for mobile users
### Enterprise Ready
- Team management features
- Admin controls and oversight
- Usage analytics and reporting
- Integration with existing tools
## Pricing
| Tier | Features | Price |
|------|----------|--------|
| :--- | :--- | :--- |
| **Free** | Free for everyone | $0 |
| **Pro** | Access our latest models<br/>Access other cloud providers without API keys | Coming Soon |
| **Enterprise** | Self-host or we host it for you<br/>Active support and SLAs<br/>SSO integration<br/>Team features | Contact Sales |
<Aside type="tip">
Free tier gives you access to Jan's core functionality. Pro and Enterprise tiers add advanced features and support.
</Aside>
## Why Use Jan Web?
### When You Need Cloud Power
- Access from any device with internet
- More powerful models than local hardware
- Faster responses for complex tasks
- Always up-to-date with latest features
### Team Collaboration
- Share knowledge and workflows
- Collaborate on projects in real-time
- Consistent experience across team
- Admin controls for organizations
### Zero Setup
- No downloads or installations
- No API keys to manage
- No complex configurations
- Just open browser and start chatting
### Enterprise Features
- SSO integration for security
- Team management and controls
- Usage analytics and reporting
- Professional support and SLAs
## For Developers
### API Access
```javascript
// Same API as local Jan
const response = await fetch('https://api.jan.ai/v1/chat/completions', {
headers: { 'Authorization': 'Bearer YOUR_KEY' },
body: JSON.stringify({
model: 'jan-nano',
messages: [{ role: 'user', content: 'Hello' }]
})
});
```
### OpenAI Compatible
Drop-in replacement for OpenAI API:
```python
# Just change the base URL
client = OpenAI(
base_url="https://api.jan.ai/v1",
api_key="your-jan-key"
)
```
### Integration Options
- REST API for custom applications
- Webhook support for automation
- Team API for organization management
- Usage analytics API for monitoring
## Common Questions
### How is this different from ChatGPT?
- Open source and transparent
- Privacy-focused by design
- Same experience as local Jan
- Team collaboration features
- Enterprise self-hosting options
### Can I use my own models?
Not directly on Jan Web. For custom models, use Jan Desktop locally or self-host Jan Server.
### Is my data secure?
Yes, Jan Web follows enterprise security practices:
- Encrypted connections and storage
- No training on user data
- Optional anonymous usage
- Regular security audits
- GDPR compliance
### When will it launch?
Beta launch is coming soon. Full public launch planned for 2025.
## Coming Soon
### Beta Launch Features
- Web-based chat interface
- Team collaboration tools
- API access for developers
- Jan Desktop/Mobile integration
- Basic admin controls
### Full Launch Features
- Advanced team management
- SSO and enterprise authentication
- Custom workflows and automation
- Usage analytics and reporting
- Advanced model access
<Aside type="note">
Jan Web is designed to feel exactly like local Jan, just accessible anywhere. Same principles, same experience, cloud convenience.
</Aside>
## The Philosophy
Cloud AI doesn't have to mean giving up control. Jan Web proves you can have:
- Convenience without surveillance
- Team collaboration without data mining
- Enterprise features without vendor lock-in
- Cloud power with privacy respect
---
[Join Beta Waitlist](https://jan.ai/beta) | [Contact Sales](https://jan.ai/enterprise) | [API Documentation](/docs/api)
| **Enterprise** | Self-host or we host it for you<br/>Active support and SLAs<br/>SSO integration<br/>Team features | Contact Sales |

209
website/src/content/products/platforms/mobile.mdx
View File

@ -1,219 +1,48 @@
---
title: Jan Mobile
description: Your AI assistant on the go
description: Your AI assistant on the go. Seamlessly connect to local, desktop, or server models.
sidebar:
order: 3
---
import { Aside, Card, CardGrid } from '@astrojs/starlight/components';
![Jan Mobile](../../../assets/jan_mobile.png)
**Status:** Coming Q4 2025
Connect to Desktop/Server, run local mode with Jan Nano or Lucy, same experience everywhere.
Jan Mobile brings the same AI experience to your phone. Connect to your desktop, your server, or run models locally.
<Aside type="note">
Coming Q4 2025. Sign up here to get a notification when it's available [Newsletter](https://jan.ai/mobile).
</Aside>
## How It Works
Jan Mobile adapts to your situation:
At Home, you can connect to your Jan Desktop over WiFi
![Jan Mobile Home](../../../assets/jan_mobile2.png)
At Work, you can connect to your company Jan Server
![Jan Mobile Home](../../../assets/jan_mobile3.png)
On the Go, you can run Jan Nano on your phone or talk to your favourite cloud-based model
![Jan Mobile Home](../../../assets/jan_mobile4.png)
**No configuration needed. It just works.**
- **At Home:** Connect to your Jan Desktop over WiFi.
- **At Work:** Connect to your company Jan Server.
- **On the Go:** Run Jan Nano on your phone or use a cloud model.
## Three Modes, One Experience
### Desktop Mode
When you're near your computer at home, you an toggle phone uses its models and processing power.
```
Your Phone → WiFi → Your Desktop → Response
(automatic) (powerful models)
```
**Benefits:**
- Access to larger models
- Faster processing
- Shared conversations
- No phone battery drain
Access larger, more powerful models running on your home computer. No phone battery drain.
### Server Mode
Connect to your organization's Jan Server for team collaboration.
```
Your Phone → Internet → Company Server → Response
(secure) (shared models)
```
**Benefits:**
- Team knowledge base
- Consistent models
- Central management
- Work anywhere
Connect to your organization's private AI cloud for team collaboration and access to shared knowledge.
### Local Mode
No connection? No problem. Jan Nano runs directly on your phone.
```
Your Phone → Jan Nano (6GB) → Response
(private & offline)
```
**Benefits:**
- Complete privacy
- Works offline
- No data usage
- Always available
<Aside type="tip">
The app automatically switches between modes based on what's available. You don't need to do anything.
</Aside>
No connection? No problem. Run models like 'Jan Nano' directly on your phone for complete privacy and offline access.
## Key Features
<CardGrid>
<Card title="Seamless Switching" icon="random">
Move from home to office to airplane. One-click and Jan adapts immediately.
</Card>
<Card title="Voice First" icon="phone">
Talk to Jan naturally. Responses can be spoken too.
</Card>
<Card title="Sync Everything" icon="up-arrow">
Conversations, settings, and preferences follow you across devices.
</Card>
<Card title="iOS and Android" icon="mobile">
Native apps for both platforms with consistent experience.
</Card>
<Card title="Three Adaptive Modes" icon="setting">
Automatically switches between Desktop, Server, and Local modes.
</Card>
<Card title="Jan Nano Integration" icon="rocket">
Run AI models directly on your phone for complete privacy.
</Card>
<Card title="Seamless Switching" icon="random" />
<Card title="Voice First Interface" icon="microphone" />
<Card title="Sync Everything" icon="sync" />
<Card title="iOS and Android" icon="mobile" />
<Card title="Adaptive Modes" icon="adjustments" />
<Card title="Jan Models Integration" icon="puzzle" />
</CardGrid>
## Privacy & Security
### Your Data, Your Control
- Local Mode: Everything stays on your phone
- Desktop Mode: Direct encrypted connection
- Server Mode: Your organization's policies apply
### No Compromises
- Biometric app lock
- Encrypted storage
- No cloud backups without permission
- Clear data anytime
## Why Mobile Matters
Your phone is with you always. Your AI assistant should be too. But that doesn't mean sacrificing
privacy or control.
Jan Mobile proves you can have:
- Powerful AI anywhere
- Complete privacy when needed
- Simple experience always
- No subscriptions or tracking
## Technical Details
### App Architecture
- Built with native performance in mind
- Small app size (~50MB download)
- Efficient battery usage
- Background processing support
### Supported Platforms
- **iOS:** iPhone and iPad support
- **Android:** Phone and tablet support
- **Voice Interface:** Natural speech input/output
- **Seamless Device Switching:** Real-time sync across devices
### Jan Nano for Mobile
- Optimized 6GB model for mobile devices
- Runs entirely on-device
- No internet required
- Complete privacy and offline functionality
## Coming Features
### Launch (Q4 2025)
- iOS and Android native apps
- Three adaptive connection modes
- Voice-first interface
- Basic chat and conversation management
- Automatic mode switching
### Post-Launch Updates
- Advanced voice input with wake words
- Background conversation sync
- Widget support for quick access
- Enhanced team collaboration features
- Custom model support
<Aside type="note">
Jan Mobile is built for native performance with efficient resource usage and small app size.
</Aside>
## The Vision
Most mobile AI apps are just cloud wrappers. Jan Mobile is different:
| Feature | Other AI Apps | Jan Mobile |
|---------|---------------|------------|
| Offline Mode | ❌ | ✅ Jan Nano |
| Desktop Connection | ❌ | ✅ Your models |
| Privacy | Your data in cloud | Your data stays yours |
| Cost | Monthly subscription | Free with your hardware |
## Development Status
### Current Progress
- Core architecture completed
- Desktop/Server connection protocols implemented
- Jan Nano mobile optimization in progress
- iOS and Android app development underway
### Beta Testing
- Closed beta planned for Q3 2025
- Open beta following successful internal testing
- Community feedback integration throughout development
## While You Wait for Jan Mobile
1. **Set up Jan Desktop** - It will power your mobile experience when at home
2. **Try Jan Web** - Get familiar with the cloud interface
3. **Join our newsletter** - Be first to know when beta launches
4. **Provide feedback** - Help shape the mobile experience through our Discord community
5. **Prepare your setup** - Ensure Jan Desktop is configured for network access
## Get Notified
Want to be among the first to try Jan Mobile?
- [Join our newsletter](https://jan.ai/mobile) for launch updates
- [Follow us on Discord](https://discord.gg/jan) for development progress
- [Check our GitHub](https://github.com/janhq/jan) for technical updates
---
[Jan Desktop](/docs/products/platforms/desktop) | [Jan Web](/docs/products/platforms/jan-ai) | [Jan Server](/docs/products/platforms/server)
- Core architecture in progress
- Desktop/Server connection protocols next
- Jan Nano mobile optimization in progress
- Beta launch planned for Q4 2025

330
website/src/content/products/platforms/server.mdx
View File

@ -1,335 +1,39 @@
---
title: Jan Server
description: Your own private AI cloud
description: Your own private AI cloud. Self-hosted AI for teams and enterprises.
sidebar:
order: 4
---
import { Aside, Card, CardGrid } from '@astrojs/starlight/components';
![Jan Server](../../../assets/jan-server.png)
**Status:** Coming Q2 2025
Self-hosted solution for teams and enterprises. Your own private AI cloud.
Jan Server is Jan Desktop with multi-user support. Deploy it on your hardware to create your own private AI cloud for your team or organization.
<Aside type="note">
Coming Q2 2025. Join the early access list at [jan.ai/server](https://jan.ai/server).
</Aside>
## What is Jan Server?
```
Jan Server = Jan Desktop + Multi-user support + Real hardware
```
It's the same engine that powers Jan Desktop, scaled up for teams. Your data stays on your servers, your models run on your GPUs, your AI remains yours.
Jan Server is a powerful AI API platform with multi-user support that you can self-host. Deploy it on your hardware to create your own private AI cloud for your team or organization, or run it at scale with Jan's managed service.
## Why Organizations Need This
### The Problem
Every API call to ChatGPT or Claude is:
- Your intellectual property leaving your network
- Potential training data for someone else's model
- A compliance nightmare waiting to happen
- A monthly bill that never ends
### The Solution
Jan Server gives you:
- **Complete control**: Your hardware, your rules
- **Total privacy**: Nothing leaves your network
- **Predictable costs**: One-time hardware investment
- **Compliance ready**: GDPR, HIPAA, SOC2 friendly
Jan Server gives you complete control over your AI infrastructure, ensuring total privacy, predictable costs, and compliance readiness.
## Key Features
<CardGrid>
<Card title="Multi-User Support" icon="users">
Support for 5-500+ concurrent users with individual accounts and permissions.
</Card>
<Card title="Enterprise Authentication" icon="shield">
SSO, LDAP integration and enterprise-grade security controls.
</Card>
<Card title="Flexible Deployment" icon="setting">
Docker, Kubernetes, or bare metal deployment options.
</Card>
<Card title="Admin Dashboard" icon="laptop">
Centralized management for users, models, and system monitoring.
</Card>
<Card title="Team Knowledge Sharing" icon="group">
Shared conversations, templates, and collaborative workflows.
</Card>
<Card title="Same API" icon="code">
Drop-in replacement for OpenAI API - no code changes needed.
</Card>
<Card title="Multi-User Support" icon="users" />
<Card title="Enterprise Authentication" icon="shield" />
<Card title="Flexible Deployment" icon="setting" />
<Card title="Admin Dashboard" icon="laptop" />
<Card title="Team Knowledge Sharing" icon="group" />
<Card title="Same API as Desktop" icon="code" />
</CardGrid>
## Deployment Options
<CardGrid>
<Card title="Docker: Single Command Setup" icon="laptop">
Perfect for getting started quickly with containerized deployment.
</Card>
<Card title="Kubernetes: Enterprise Scale" icon="building">
Auto-scaling, high availability, and enterprise orchestration.
</Card>
<Card title="Bare Metal: Maximum Control" icon="setting">
Direct hardware access for maximum performance and customization.
</Card>
</CardGrid>
## Simple Deployment
### Docker (Single Command Setup)
```yaml
version: '3'
services:
jan-server:
image: jan.ai/server:latest
ports:
- "80:80"
- "1337:1337"
volumes:
- ./models:/models
- ./data:/data
deploy:
resources:
reservations:
devices:
- capabilities: [gpu]
```
### Kubernetes (Enterprise Scale)
```yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: jan-server
spec:
replicas: 3
selector:
matchLabels:
app: jan-server
template:
metadata:
labels:
app: jan-server
spec:
containers:
- name: jan-server
image: jan.ai/server:latest
resources:
limits:
nvidia.com/gpu: 1
```
### Bare Metal (Maximum Control)
Direct installation on your hardware for maximum performance and custom configurations.
## Detailed Features
### Enterprise Authentication
- **SSO Integration**: SAML, OAuth, OpenID Connect
- **LDAP/Active Directory**: Existing user directory integration
- **Role-Based Access**: Granular permissions and model access
- **API Key Management**: Individual and service account keys
### Multi-User Management
- **User Accounts**: Individual profiles and preferences
- **Usage Tracking**: Per-user analytics and quotas
- **Team Collaboration**: Shared conversations and workflows
- **Admin Dashboard**: Centralized user and system management
### Same API as Desktop
```python
# Your code doesn't change
client = OpenAI(
base_url="https://jan.company.internal/v1",
api_key="user-specific-key"
)
```
### Team Knowledge Sharing
- **Shared Conversations**: Collaborative chat threads
- **Template Library**: Reusable prompts and workflows
- **Knowledge Base**: Organizational AI knowledge
- **Team Analytics**: Usage patterns and insights
<Aside type="tip">
Jan Server uses the same models as Desktop. No special "enterprise" versions with inflated prices.
</Aside>
- **Docker:** Single command setup
- **Kubernetes:** Scale with your needs
- **Bare Metal:** Maximum control and performance
## Scaling Guidelines
### Small Teams (5-10 users)
- **Hardware**: Single RTX 6000 Ada (48GB VRAM)
- **RAM**: 128GB system memory
- **Models**: Up to 70B parameter models
- **Concurrent Users**: 5-10 active users
- **Estimated Cost**: ~$15,000 one-time hardware
### Department Scale (10-50 users)
- **Hardware**: 2-4 GPU cluster nodes
- **RAM**: 256GB per node
- **Models**: Multiple concurrent model instances
- **Concurrent Users**: 10-50 active users
- **Estimated Cost**: ~$50,000-$100,000 one-time
### Enterprise Scale (50+ users)
- **Hardware**: DGX cluster or custom configuration
- **RAM**: Scalable as needed
- **Models**: Full model library with redundancy
- **Concurrent Users**: 50-500+ active users
- **Estimated Cost**: Custom enterprise quote
## Real-World Use Cases
| Organization Type | Deployment | Benefits Achieved |
|------------------|------------|-------------------|
| Law Firm | 2x RTX 6000, 200 users | Client data never leaves network, GDPR compliance |
| Hospital System | DGX node, 500 users | HIPAA compliant AI assistant, medical data privacy |
| Tech Startup | 4x RTX 4090, 50 users | 90% cost reduction vs. OpenAI API calls |
| Research University | Multi-node cluster | Unrestricted research, no usage limits |
| Financial Services | Air-gapped deployment | Complete data isolation, regulatory compliance |
## Hardware Requirements
### Minimum Configuration
- **GPU**: RTX 3090 or better (24GB VRAM minimum)
- **CPU**: 16+ cores (Xeon, EPYC, or equivalent)
- **RAM**: 64GB system memory minimum
- **Storage**: 1TB NVMe SSD for models and data
- **Network**: Gigabit Ethernet minimum
### Recommended Production Setup
- **GPU**: RTX 6000 Ada (48GB) or A100 (80GB)
- **CPU**: Dual socket Xeon/EPYC (32+ cores)
- **RAM**: 128-256GB system memory
- **Storage**: RAID NVMe array (2TB+ capacity)
- **Network**: 10Gbps for multiple concurrent users
### Enterprise Scaling Guidelines
- **Users per GPU**: ~5-10 concurrent active users
- **70B Models**: Require 40-80GB VRAM depending on quantization
- **CPU Fallback**: Smaller models can run on CPU for cost optimization
- **High Availability**: Multi-node deployment with load balancing
- **Backup Strategy**: Regular model and data backups recommended
## Why Self-Host?
### For IT Teams
- No data leaves your network
- Complete audit trails
- Integrate with existing auth (LDAP/AD)
- Predictable resource usage
### For Security Teams
- Air-gapped deployment options
- End-to-end encryption
- No third-party access
- Full compliance control
### For Finance Teams
- One-time hardware cost
- No per-token pricing
- Predictable TCO
- Use existing infrastructure
## Development Roadmap
### Q2 2025 Launch Features
- Multi-user authentication and management
- Web-based admin dashboard
- OpenAI API compatibility
- Docker and Kubernetes deployment
- Basic usage monitoring and analytics
- Team collaboration features
### Post-Launch Updates
- Advanced governance and compliance tools
- Model fine-tuning interface
- Automated scaling and load balancing
- Comprehensive backup and restore
- Enhanced security and audit logging
### Future Vision
- Federated multi-site deployments
- Cross-region synchronization
- Advanced usage analytics and insights
- Custom model training and optimization
- Integration with enterprise workflow tools
<Aside type="caution">
Jan Server requires proper cooling and power for GPU hardware. Plan your infrastructure accordingly.
</Aside>
## Migration Path
### From Cloud AI
1. Deploy Jan Server
2. Import your workflows
3. Update API endpoints
4. Migrate users gradually
### From Jan Desktop
1. Same models work instantly
2. Add user management
3. Scale as needed
## The Philosophy
We believe organizations should own their AI infrastructure just like they own their data. Jan Server makes this possible without compromising on capabilities.
This isn't about avoiding the cloud - it's about having a choice. Run your AI where it makes sense for your organization.
## Support Options
### Community Edition
- Full features
- Community support
- Perfect for small teams
### Enterprise Edition
- Priority support
- Custom deployment help
- SLA guarantees
- Training included
## Get Ready for Jan Server
While Jan Server is in development (Q2 2025 launch):
### Preparation Steps
1. **Assess Hardware Needs**: Review our scaling guidelines above
2. **Plan Network Architecture**: Consider security and access requirements
3. **Evaluate Authentication**: Determine SSO/LDAP integration needs
4. **Test with Jan Desktop**: Same models and API, perfect for preparation
5. **Join Early Access Program**: Get notified when beta testing begins
### Early Access Benefits
- Beta testing opportunities
- Hardware optimization guidance
- Deployment planning assistance
- Priority technical support
- Input on feature development
### Enterprise Consultation
For large deployments or custom requirements:
- Architecture planning sessions
- Hardware recommendation consultations
- Security and compliance reviews
- Migration planning from existing AI services
- Custom deployment assistance
---
**Ready to Plan Your Deployment?**
[Join Early Access](https://jan.ai/server) | [Enterprise Consultation](https://jan.ai/enterprise) | [Hardware Planning Guide](/docs/server/hardware) | [Jan Desktop Trial](/download)
- **Small Teams (5-10 users):** Small GPU cluster
- **Departments (10-50 users):** 4-8 GPU cluster nodes
- **Enterprise (50+ users):** Multi-cluster setup with custom configurations

30
website/src/content/products/tools/browseruse.mdx
View File

@ -3,10 +3,12 @@ title: BrowserUse
description: Native browser automation for Jan, enabling AI to interact with the web on your behalf.
sidebar:
order: 3
banner:
content: 'In Development: This tool is planned and not yet available.'
---
import { Aside, Card, CardGrid } from '@astrojs-starlight/components';
import { Aside } from '@astrojs/starlight/components';
<Aside type="caution" title="In Development">
This tool is planned and not yet available.
</Aside>
## Let Jan Use Your Browser
@ -14,23 +16,11 @@ import { Aside, Card, CardGrid } from '@astrojs-starlight/components';
Think of it as an integrated, privacy-first automation layer that turns Jan from a conversational AI into a powerful agent for action.
<Aside type="note">
**This tool is not yet available.** We are designing `BrowserUse` to be a core component of Jan's agentic capabilities.
</Aside>
### Inspired by the Best
While tools like [Browser Use](https://docs.browser-use.com/introduction) exist for developers, Jan's `BrowserUse` will be a built-in, user-friendly feature. No complex setup required—just grant permission and let Jan handle the rest.
### Built on MCP
The tool will be implemented as a native **Model Context Protocol (MCP)** server within Jan, ensuring secure and standardized communication between the AI model and the browser. This allows for powerful, auditable, and secure web interactions.
---
The tool will be implemented as a native **Model Context Protocol (MCP)** server within Jan, ensuring secure and standardized communication between the AI model and the browser.
### Planned Core Features:
- **Secure Sessions**: All browsing activity happens in an isolated, sandboxed environment.
- **Natural Language Control**: Simply tell Jan what you want to do (e.g., "Book a flight," "Research this topic and summarize the top 5 findings," "Apply for this job for me").
- **Visual Understanding**: Jan will be able to see and interpret the content on pages, not just the underlying code.
- **User in the Loop**: Always ask for permission before taking critical actions like submitting forms or making purchases.
- **Secure Sessions** in an isolated, sandboxed environment.
- **Natural Language Control** (e.g., 'Book a flight...')
- **Visual Understanding** to interpret page content.
- **User in the Loop** for critical actions.

30
website/src/content/products/tools/deepresearch.mdx
View File

@ -1,31 +1,27 @@
---
title: DeepResearch
description: Go beyond simple search with an AI agent that performs comprehensive, multi-step research for you.
description: An AI agent that performs comprehensive, multi-step research for you.
sidebar:
order: 1
banner:
content: 'In Development: This tool is planned and not yet available.'
---
import { Aside, Card, CardGrid } from '@astrojs/starlight/components';
import { Aside } from '@astrojs/starlight/components';
<Aside type="caution" title="In Development">
This tool is planned and not yet available.
</Aside>
## Your Personal Research Analyst
`DeepResearch` is a planned native tool for Jan that transforms it into a powerful research agent. Give it a complex question, and it will autonomously browse, analyze, and synthesize information from numerous sources to deliver a comprehensive, structured report.
Think of it as Jan's answer to the advanced research capabilities seen in [OpenAI's ChatGPT](https://openai.com/index/introducing-deep-research/) and [Google's Gemini](https://gemini.google/overview/deep-research/), but built with privacy and user control at its core.
<Aside type="note">
**This tool is not yet available.** We are building `DeepResearch` to handle complex queries that would normally take a human hours of work.
</Aside>
Think of it as Jan's answer to the advanced research capabilities seen in **OpenAI's ChatGPT** and **Google's Gemini**, but built in the open and with user control at its core.
### How It Will Work
Unlike a simple web search that returns a list of links, `DeepResearch` will understand your goal, create a research plan, execute it, and deliver a final, synthesized document with citations.
Unlike a simple web search that returns a list of links, `DeepResearch` will understand your goal, create a research plan that you can edit, execute it, and deliver a final, synthesized document with citations.
### Planned Core Features:
- **Autonomous Multi-Step Research**: Deconstructs complex questions into logical steps and executes them.
- **Comprehensive Source Analysis**: Can read and synthesize information from web pages, PDFs, and other documents.
- **Structured Report Generation**: Delivers well-organized reports, not just chat messages.
- **Full Transparency**: Cites all sources and shows its work, so you can verify the information.
- **Local-First Privacy**: Conduct research using local models to ensure your queries and findings remain private.
- **Autonomous Multi-Step Research**
- **Comprehensive Source Analysis**
- **Structured Report Generation**
- **Full Transparency with Citations**
- **Local-First Privacy**

35
website/src/content/products/tools/search.mdx
View File

@ -3,32 +3,27 @@ title: Search
description: A native search tool that gives you answers, not just links, with complete privacy.
sidebar:
order: 2
banner:
content: 'In Development: This tool is planned and not yet available.'
---
import { Aside, Card, CardGrid } from '@astrojs/starlight/components';
import { Aside } from '@astrojs/starlight/components';
<Aside type="caution" title="In Development">
This tool is planned and not yet available.
</Aside>
## Answers, Not Just Links
`Search` is a planned native tool for Jan that rethinks web search. Instead of just giving you a list of links to sift through, it understands your question, scours the web, and provides a direct, synthesized answer with sources cited.
'Search' is a planned native tool for Jan that rethinks web search. Instead of just giving you a list of links to sift through, it understands your question, scours the web, and provides a direct, synthesized answer with sources cited.
Think of it as a private, self-hosted alternative to services like Perplexity.ai, integrated directly into your AI assistant.
<Aside type="note">
**This tool is not yet available.** We are building `Search` to be the default way Jan accesses real-time information from the web.
</Aside>
Think of it as a private, self-hosted alternative to services like **Perplexity.ai**, integrated directly into your AI assistant.
### How It's Different
- **Privacy-First**: Your search queries are processed locally and anonymized. What you search for is your business.
- **Direct Answers**: Get a concise, accurate answer compiled from the best sources, not just ten blue links.
- **Cited Sources**: Every piece of information is backed by a verifiable source, so you can trust the answer or dig deeper.
- **Conversational Follow-up**: Ask follow-up questions in a natural way, and Jan will maintain context.
- **Privacy-First:** Your search queries are processed locally and anonymized.
- **Direct Answers:** Get a concise, accurate answer compiled from the best sources.
- **Cited Sources:** Every piece of information is backed by a verifiable source.
- **Conversational Follow-up:** Ask follow-up questions in a natural way.
### Planned Core Features:
- **Real-Time Information**: Access up-to-the-minute news, data, and events.
- **Source Verification**: Prioritizes authoritative and reliable sources.
- **Customizable Focus**: Tailor your search to specific domains like academic papers, news, or code repositories.
- **Seamless Integration**: Works hand-in-hand with other tools like `DeepResearch` and `BrowserUse` for more complex tasks.
---
- **Real-Time Information**
- **Source Verification**
- **Customizable Focus**
- **Seamless Integration** with other tools

2
website/src/pages/products/index.astro
View File

@ -6,7 +6,7 @@ import { Content } from '../../content/products/index.mdx';
<StarlightPage
frontmatter={{
title: 'Products',
title: 'Vision 🚀',
description: 'AI that runs where you need it, how you need it',
tableOfContents: false
}}

4
website/src/pages/products/tools/browseruse.astro
View File

@ -6,8 +6,8 @@ import { Content } from '../../../content/products/tools/browseruse.mdx';
<StarlightPage
frontmatter={{
title: 'BrowserUse',
description: 'AI-powered browser automation and web interaction tool',
template: 'splash'
description: 'AI-powered browser automation and web interaction tool',
tableOfContents: false
}}
sidebar={[
{