Skip to content

Open Superintelligence Platform

“Superintelligence isn’t one massive model. It’s an ecosystem of specialized models, tools, and applications working together - and you should own all of it.” — Jan Philosophy

Open superintelligence is AI that adapts to how you want to work, not the other way around. It’s not about forcing you to choose between local or cloud, powerful or private, easy or extensible. It’s about having it all, and owning it.

We’re not trying to build GPT-5. We’re building something better: specialized models that excel at specific tasks, tools that extend their capabilities, and applications that tie it all together.

Models (specialized for tasks)
+ Tools (extend capabilities)
+ Applications (work everywhere)
= Superintelligence you own

The “one model to rule them all” approach has fundamental flaws:

  • Inefficient: Using the same weights for poetry and mathematics
  • Inflexible: Can’t optimize for specific use cases
  • Expensive: Massive compute for simple tasks
  • Monolithic: Single point of failure and control
  • Jan-Search: Knows how to find and synthesize information
  • Jan-Write: Understands structure, tone, and creativity
  • Jan-Analyze: Excels at reasoning and data interpretation
  • Jan-Code: Optimized for programming tasks

Each model does one thing brilliantly. Together, they form superintelligence.

Not just our models, but any model you need:

  • Jan Models: Purpose-built for specific tasks
  • Community Models: Any GGUF from Hugging Face
  • Cloud Models: Connect to OpenAI, Anthropic when needed
  • Your Models: Fine-tune and deploy your own

Models need capabilities beyond conversation:

  • Search: Get answers, not just links
  • DeepResearch: Autonomous multi-step investigation
  • BrowserUse: Let AI interact with the web
  • MCP Protocol: Connect any data source or API

Same experience, infinite deployment options:

  • Jan Desktop: Your local AI workstation
  • Jan.ai: Browser-based, no setup required
  • Jan Mobile: AI that follows you (coming soon)
  • Jan Server: Self-hosted for teams (coming soon)

True ownership means having choices:

  • Full Local: Everything on your hardware, works offline
  • Full Cloud: We host it, you just use it
  • Hybrid: Local for sensitive work, cloud for heavy compute
  • Self-Hosted: Your servers, your control
  • Use our models when they’re best
  • Use community models when they fit
  • Use cloud models when you need them
  • Train your own when you want to
  • Personal: Run on your laptop
  • Team: Deploy on your server
  • Enterprise: Scale across infrastructure
  • Global: Distribute to the edge

We don’t just open source our code. We open source our entire development process:

Live at train.jan.ai:

  • See models training in real-time
  • View our datasets and methods
  • Learn from our failures
  • Track actual progress

Evaluate at eval.jan.ai:

  • Compare model outputs
  • Vote on what works
  • Suggest improvements
  • Access all evaluation data

Everything on GitHub:

  • Contribute features
  • Report issues
  • Fork and modify
  • Join discussions

Open systems create compound benefits:

  • More models to choose from
  • More tools to extend capabilities
  • More deployment options
  • More control over your AI
  • Build on a stable platform
  • Access to growing ecosystem
  • No vendor lock-in fears
  • Community support
  • Deploy how you need
  • Customize for your use case
  • Control your costs
  • Own your infrastructure

Jan Exam ensures excellence across the ecosystem:

  • Objective benchmarks
  • Real-world testing
  • Community validation
  • Transparent results

Whether it’s our model or someone else’s, if it performs well, it belongs on Jan.

  • Jan Desktop with local and cloud models
  • Basic tools via MCP
  • Growing model ecosystem
  • Active community
  • Jan v1 specialized models
  • Advanced tools (search, research, browser)
  • Jan Server for teams
  • Mobile applications
  • Models that understand your context
  • Tools that act autonomously
  • Applications that work everywhere
  • An ecosystem owned by its users
  • No vendor lock-in vs. total dependence
  • Own forever vs. rent forever
  • Infinite flexibility vs. their way only
  • Community innovation vs. corporate roadmap
  • Complete ecosystem vs. fragmented tools
  • Works out of box vs. endless configuration
  • Unified experience vs. duct-tape integration
  • Professional polish vs. research prototypes

This isn’t just about building better AI. It’s about ensuring AI serves humanity, not the other way around.

When you use Jan, you’re not just a user. You’re part of a movement building:

  • AI that works anywhere
  • Intelligence you can own
  • Tools that adapt to you
  • A future without gatekeepers

We promise to build superintelligence that:

  1. Works Everywhere: From laptop to data center to edge
  2. Belongs to You: Download it, modify it, own it forever
  3. Stays Open: Core will always be open source
  4. Keeps Improving: Community-driven development
  5. Respects Choice: Your deployment, your rules

Superintelligence shouldn’t be controlled by a few companies. It shouldn’t force you to choose between power and ownership. It shouldn’t lock you into one way of working.

With Jan, it doesn’t.

You get superintelligence that adapts to how you work, runs where you need it, and belongs to you completely.

That’s open superintelligence. That’s what we’re building.


“The best AI platform is one where choosing local or cloud is like choosing between laptop or desktop - same experience, different form factor, your choice.”