112 lines
4.4 KiB
Plaintext
112 lines
4.4 KiB
Plaintext
---
|
|
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';
|
|
|
|
|
|
> **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 through a combination of
|
|
models, applications and tools. For this we are...
|
|
|
|
> **building Jan as the ecosystem that ties all of these seamlessly so that users, regardless of their technical
|
|
background, add intelligence to their day-to-day lives like they would, but better, with similar tools.**
|
|
|
|
|
|
|
|
|
|
## Core Principles
|
|
|
|
### 1) Build the Full Stack
|
|
|
|
Models alone aren't enough. Neither are applications. Superintelligence requires models that
|
|
understand your needs, tools that extend capabilities, and applications that tie it all
|
|
together. We're building all three, openly.
|
|
|
|
### 2) You Choose Who Runs It
|
|
Run Jan on your laptop. Self-host it on your servers. Use our cloud. The same superintelligence
|
|
works everywhere. Your data, your compute, your choice.
|
|
|
|
### 3) Start Simple, Scale Infinitely
|
|
Open Jan and start chatting. No setup required. When you need more - better models, advanced tools,
|
|
team deployment - everything's there. The complexity scales with your ambition, not our architecture.
|
|
|
|
## The Path to Superintelligence
|
|
|
|
### Today we have the Foundation
|
|
|
|
- **A desktop app** that works both with local and cloud-based models
|
|
- **Jan models** small enough to run on any laptop and powerful enough to scale on any server
|
|
- **Basic tools** enabled through MCP Search, file parsing, simple workflows
|
|
|
|
### Next 12 Months: Ecosystem
|
|
- **The Jan v1 models** are a specialized series of models with general capabilities but optimsed
|
|
for specific tasks like search, analysis, creative writing and more
|
|
- **The Jan server** works as a self-hosted AI infrastructure for teams
|
|
- **Advanced tools** like browser use, deep research, and long-term memory works across devices, excels
|
|
across different day-to-day use cases, and scales with the needs of large teams
|
|
- **Cross-device sync** allows you to take your AI everywhere
|
|
|
|
### End State: Open Superintelligence
|
|
Not one massive model, but an ecosystem of specialized models, tools, and applications working
|
|
together. Built in public. Owned by whoever runs it.
|
|
|
|
## Why This Matters: The Status Quo
|
|
|
|
Every other AI company wants to be your AI provider. We want you to own your AI.
|
|
|
|
- **OpenAI/Anthropic**: Their models, their servers, their rules
|
|
- **Open Source Models**: Powerful but fragmented - no cohesive experience
|
|
- **Jan**: Complete ecosystem you can own, modify, and deploy however you want
|
|
|
|
## Watch Us Build
|
|
|
|
### Live Model Training
|
|
We train our models in public. Check the [models page](./models/jan-v1) to see:
|
|
- Real-time training progress
|
|
- Failed runs and what went wrong
|
|
- Models in testing before release
|
|
|
|
No "trust us, it's good." Watch the entire process from dataset to deployment.
|
|
|
|
### Help Evaluate Our Models
|
|
Every model needs real-world testing. Join our open evaluation platform where you can:
|
|
- Compare model outputs side-by-side
|
|
- Test specific capabilities you care about
|
|
- Vote on which responses actually help
|
|
- Suggest improvements based on your use cases
|
|
|
|
Think LMArena, but you can see all the data, run your own evals, and directly influence what we train next.
|
|
|
|
[Test/evaluate our models here](link)
|
|
|
|
### Our Models Training Right Now
|
|
|
|
We don't just talk about open development. Here's what's actually happening:
|
|
|
|
| Model | Progress | Status | Details |
|
|
|:------|:---------|:-------|:--------|
|
|
| **Jan-Search-7B** | ████████░░ 82% | Testing | [View run](/) • 2.1M steps • ETA 3 days |
|
|
| **Jan-Write-13B** | ████░░░░░░ 41% | Training | [View run](/) • 980K steps • On track |
|
|
| **Jan-Analyze-13B** | ████████░░ ~~67%~~ | Failed | [View logs](/) • OOM at step 1.5M • Restarting |
|
|
|
|
These are our actual models training on our hardware in our Singapore office. Click any run to see:
|
|
- Live loss curves
|
|
- Training datasets
|
|
- Evaluation metrics
|
|
- Even our failures
|
|
|
|
[Watch live training →](/train)
|
|
|
|
|
|
## Get Involved
|
|
|
|
We build in public. Everything from our model training to our product roadmap is open.
|
|
|
|
- [GitHub](link) - Contribute code
|
|
- [Handbook](link) - See how we train models
|
|
- [Discord](link) - Join the discussion
|