Catamaran AI

Private access to large language models running on our own hardware. Create an account and you get an API key immediately — no waitlist, no card.

Create your account I already have an account

Running right now

Qwen 3.8 · 27B128K context · Vision — reads images and long video
Muse Glimmer · 30B128K context · Long context, tool-calling, text only
Qwen 3.6 · 35B256K context · General purpose, tool-calling, very long context

These are the models serving at this moment. Your key works with whatever is up — and there is a lot more in the full catalog, including video generation and quantum simulation.

Want your own private AI?

Everything here runs on hardware we own and rack ourselves. If you want a private machine — your models, your data, nobody else on the GPUs — we have rack space and we do it soup to nuts. Tell us what you are trying to do and we will tell you honestly whether it makes sense.

Tip jar

Access is free. If it is useful to you, tips keep the GPUs fed — entirely optional. Scan with your phone, or tap.

We like to hear from you

Tell us your thoughts — what worked, what did not, what you wish existed. Your feedback is appreciated and it genuinely shapes what we host next.