Skip to content

Official documentation

Take back control of your generative AI usage

Xolo sits between your applications and the model providers. It holds the API keys, the budgets and the routing between models, and keeps a record of every request. One binary, licensed under AGPL-3.0.

Install Xolo Understand the model

LLM CLIENTS PROVIDERS Chat OpenAI SDK Agents tools, RAG IDE / CLI application key Remote API OpenAI, Azure… European API Mistral, OVH… Self-hosted Ollama, vLLM, others… Xolo 1 · AUTHENTICATE Application key role and permissions 2 · CONTROL Budget and quota organisation, member 3 · ROUTE Virtual model pipeline: plugins, models fallback and weighting One event per request cost · latency · energy

Where to start

Three tracks, depending on your role

In five minutes

A local instance, a key, a first call.

$ docker run -p 8080:8080 \
    ghcr.io/xolo-gateway/xolo:latest

# then, once the organisation is created
$ curl localhost:8080/v1/chat/completions \
    -H "Authorization: Bearer $XOLO_KEY" \
    -d '{"model":"gpt-4o-mini","messages":[…]}'

$ git clone https://github.com/xolo-gateway/xolo
$ cd xolo && make run

# the interface listens on http://localhost:8080