Open source
Private AI workspace
Libre WebUI
A focused workspace for your models, conversations, and ideas. Run local AI on your own terms.
- Local models
- No telemetry
- Self-hosted
Independent lab. Open possibilities.
Powerful AI.
On your terms.
Tools you can run. Models you can own. We build AI products that give you more capability, without giving up control.
Your infrastructure. Your data. Your choice.
Meet the Libre Stack01—04
Libre WebUI
Open sourceA focused workspace for your models, conversations, and ideas. Run local AI on your own terms.
Explore Libre WebUI (opens in a new tab) Open tools. Connected ideas. One independent lab.
01Private by design
Keep control of what matters.
02Open where it matters
Inspect, adapt, and make it yours.
03Built without lock-in
Choose your own way forward.
The Libre Stack
Good tools. Real independence.
Start with what you need. Run it yourself or choose a hosted product.
Hosted · Free tier
Answers from your knowledge
Libre Bot
Turn your documentation into helpful answers on your website. Your content, your voice, your assistant.
- Your sources
- Website embed
- Ready to use
Open source
An agent under your control
Libre Claw
Bring models, tools, and local memory together in an agent runtime you can inspect and run yourself.
- Local memory
- Permissioned tools
- Multi-provider
Early access
A simpler business phone
Libre Phone
Calls, texts, voicemail, and contacts in one lightweight phone desk, powered by your own infrastructure.
- Voice & SMS
- Voicemail
- Your number
Building something more specific? We can help shape the system around your team.
Let’s talkWork with the lab
Your hardest AI problem.
Our next working session.
From the first architecture decision to a system your team can own. Work directly with the people building it.
01 / Product & infrastructure
Build your AI product
Private interfaces, useful agents, and infrastructure ready for the real world.
Explore product development02 / Models & datasets
Make the model yours
Training, fine-tuning, and evaluation around the behavior your product needs.
Explore model engineering03 / Safety & resilience
Know where it breaks
Test the prompts, tools, and permissions your AI system depends on.
Explore evaluationResearch, out in the open
We build it.
We break it.
We share what we learn.
Genesis 1B is our foundation model, trained from scratch on two consumer GPUs. Follow the experiments, breakthroughs, and honest postmortems.
Inside the lab notebook1Bparameters
2× 4090consumer GPUs
80,000Run 2 steps
Make it yours
Let’s build something
you can own.
Tell us what you’re working on. We’ll find the right place to start.