How it works

cLLMHub is a hosted gateway for self-hosted LLMs. Publish a model from your own hardware, get an API key, and call it from anywhere through an OpenAI-compatible endpoint.

Three steps to your first request

01

Publish your model

Install the CLI on the machine running Ollama, vLLM, llama.cpp, MLX, or any OpenAI-compatible server. Run cllmhub publish to register the model. The CLI keeps an outbound WebSocket open — no inbound ports.

02

Create an API key

Generate a key from the dashboard. Optionally scope it to specific models you have published or cap its daily request count.

03

Call your model

Point any OpenAI SDK at api.cllmhub.com/v1 with the model name you published. Requests stream through the gateway to your hardware and back. Streaming, tool calls, and vision all work.

Your weights stay home

Inference runs on your hardware. The CLI opens an outbound WebSocket from your machine to api.cllmhub.com; the gateway forwards each request and streams the response back. Model weights are never uploaded, and your machine never has to accept an inbound connection.

If your machine goes offline, the model returns offline to callers. When it's back, it's back — no re-publish needed.

Simple monthly pricing

Free for hobby use. Pro at $3/month for 20,000 requests/day and 10 models. Max at $20/month for unlimited everything. Billed via PayPal, cancel anytime.

See plans