modelhaus gives you a single key and invoice for frontier open models we host on Canadian B300 GPUs - plus the best external models, aggregated. OpenAI-compatible, drop-in.
# OpenAI-compatible - just change the base URL curl https://modelhaus.ai/v1/chat/completions \ -H "Authorization: Bearer $MODELHAUS_KEY" \ -d '{"model":"deepseek-v4-pro","messages":[{"role":"user","content":"Hello"}]}'
Hosted sovereign on our hardware, or aggregated from the best providers - one key for all.
Built for developers who want frontier capability without the lock-in.
Call our sovereign-hosted models and external providers through a single OpenAI-compatible endpoint and key.
Our flagship open models run on Canadian B300 GPUs - your data stays in-country, not on a US hyperscaler.
Owned hardware means we undercut on the open models - frontier quality at a fraction of closed-model prices.
Works with the OpenAI SDK, LangChain, and your existing tooling. Change one base URL and go.
Per-request token accounting, live model status, and usage dashboards. No surprises.
Native tool-calling and 1M-token context on our flagship models for real agentic workloads.
Tell us what you want to run - open weights, a fine-tune, or a specific provider - and we'll stand it up. Dedicated and private deployments available.
Request a model