ai inference

Open models, OpenAI-compatible.

Point your existing OpenAI SDK at RumptyCloud and run open models on GPUs in Lagos — with streaming on by default and your prompts kept out of anyone's training data.

ai inference · new

OpenAI compatible.
Swap your base_url. Done.

Run open models on our GPUs in Lagos behind an OpenAI-compatible API. Keep the SDK you already use — just point it at RumptyCloud and ship.

  • Drop-inWorks with the official OpenAI SDKs and anything that speaks the OpenAI API. No rewrites, no new client.
  • StreamingToken-by-token responses over SSE, on by default — same shapes you already handle.
  • YoursYour prompts aren't training data. Inference runs on our own hardware, powered by the sun.
base urlhttps://chat.rumptycloud.com/v1