OGX

Connect an application

OGX, formerly Llama Stack, supports chat and document search through OpenAI-compatible clients. It runs as a server with replaceable backends, including Ollama and vLLM.

Start with the model list

For a first local check, follow the setup guide: install uv and Ollama, pull a model, then start OGX. Before writing application code, request the registered-model list. An empty list or failed request leads to the guide’s troubleshooting section. This is a documented setup check, not our test result.

The guide mixes HTTP examples with an HTTPS default. Suggested check: match the client address to your server configuration.