PhenixCode Assistant
A self-hosted, privacy-focused retrieval-augmented assistant, written in C++, that indexes a codebase locally and answers questions about it using local or cloud language models instead of a paid, cloud-only coding assistant subscription.
PhenixCode Assistant indexes a codebase locally, using local embeddings and fast vector search through Hnswlib with metadata stored in SQLite, and answers questions about the code with context-aware, chat-based responses rather than inline autocomplete. It is built for developers who want their code to stay on their own machine, want to avoid a recurring subscription by running entirely on local models, or want the flexibility to switch between local and cloud language models as needed.
It exposes both a command-line interface and an HTTP API with search, chat, and embedding endpoints, along with a metrics endpoint in JSON and Prometheus formats and health checks. Security features include JWT authentication, password-protected admin endpoints, and hashed password storage, and the project ships console and web setup wizards, service-installation scripts for Windows, Linux, and macOS, and prebuilt binaries as well as source.