One API key.Every LLM.
Donate spare API keys. Get a single OpenAI-compatible endpoint that round-robins across all of them.
$ cat problem.txt
> 6 vendors. 6 keys. 6 billing dashboards.
> Rate limits hit at 2am. No fallback.
> Half your code is API key management.
$ gpupoor solve
> Donate keys → Pool rotates → One endpoint works
$ gpupoor --help
How it works
Drop in an API key from any vendor. We validate it's alive.
Keys enter the shared pool. Health-checked every 10 minutes.
Hit one endpoint. We round-robin across healthy keys.
OpenAI-compatible. Drop-in replacement. Zero config.
Everything you need
We handle the infrastructure so you can focus on building.
One standard format for all models. No more vendor-specific adapters.
We automatically switch keys when limits are hit or errors occur.
Authenticated donations and rate limiting to protect donor keys.
Smart routing to keys with remaining quota, maximizing availability.
Works with OpenRouter, OpenAI, Anthropic, Gemini, and more.
Built by and for the GPU-poor community to democratize access.
$ gpupoor vendors --list
FAQ
Ready to donate?
Help someone ship their side project tonight.
or press [d] to go to dashboard