{"id":"llama-swap","name":"llama-swap","maker":"mostlygeek","url":"https://github.com/mostlygeek/llama-swap","repo":"mostlygeek/llama-swap","summary":"A proxy that starts and swaps local model servers on demand behind one OpenAI-compatible endpoint.","details":"llama-swap is a single Go binary configured by file, which launches llama.cpp, vLLM or another server for the requested model and stops idle ones. It runs locally.","open_source":true,"hosted":false,"speaks":["openai-api"],"also":["gateways"],"verified":"2026-11-15","category":"runtimes","category_name":"Local and self-hosted model runtimes","page":"https://modelmoot.com/tools/llama-swap","reviewed":"2026-11-15","checked":{"at":"2026-11-15T14:03:44.997Z","address_answers":true,"repository":{"at":"2026-11-16T02:35:24.378Z","archived":false,"licence":"MIT","language":"Go","last_change":"2026-11-15T21:29:41.000Z","latest_release":{"tag":"v273","published":"2026-11-15T06:17:31.000Z"}}},"note":"Descriptions are the site's own. Facts under \"checked\" were read automatically from the tool's address and public repository."}