Launched this week

Cerebellum Cloud
Run your custom models, pay per token
4 followers
Run your custom models, pay per token
4 followers
Cerebellum Cloud hosts your custom models with per-token pricing. Bring full weights or a LoRA fine-tune, get an endpoint you can call in minutes, and pay nothing while your model is idle.








Free Options
Launch Team

AWS Startup AdvisorPersonalized AWS guidance that meets you where you’re already building
Promoted
Hey Product Hunt! We're Sahil and Suhaas, and we built Cerebellum Cloud.
If you’ve fine-tuned your own model, you know the work doesn’t end when training does. Hosting it is a whole problem in itself. With Cerebellum Cloud, bring full weights or a LoRA fine-tune, and get an endpoint in minutes. Pricing is per token, and idle time costs nothing.
We built our inference stack for custom models, with fast cold starts, multi-LoRA serving, model swapping and smart GPU routing, so low or bursty traffic still gets great performance at an affordable cost.
Every new account gets $25 in free credits.
We'd love your feedback. What models are you running, and where do you host them today?