This index points to documentation for the current local developer-workstation version of NPU Proxy.
- README.md - project overview, quick start, and local
127.0.0.1:8080usage - SPEC.md - current implementation snapshot and release-truth behavior
- CHANGELOG.md - released changes
- api/OLLAMA_API.md - Ollama-compatible endpoints, request/response shapes, and local-client notes
- api/EMBEDDINGS.md - OpenAI/Ollama embedding endpoints, validation limits, and response behavior
- api/STREAMING.md - OpenAI SSE and Ollama NDJSON streaming formats and cancellation notes
- guides/MODEL_DOWNLOAD.md - Hugging Face/OpenVINO model download and conversion workflows
- guides/BENCHMARKS.md - benchmark helper usage for local performance checks
- development/CONVERTER_API.md - model converter API behavior and integration notes
- research/ - historical background material and exploratory notes; useful context, but not release commitments or current product behavior
For shipped behavior and current validation truth, prefer the root README.md, SPEC.md, and the API/guide docs above.