Ragpi

Ragpi is a self-hosted AI assistant API for teams that want to answer technical questions using their own project material. It lets developers add documentation-based support to a product, website or community.

• Knowledge sources: Build knowledge bases from documentation websites, GitHub issues, READMEs, GitHub PDFs and REST API responses, with automatic document synchronization after adding a source.
• Document search: Combine keyword and semantic search to find exact technical terms and related passages, while the assistant retrieves relevant documents as it works through a question.
• Local or cloud models: Use Ollama for response generation and embeddings, or connect OpenAI, Deepseek and OpenAI-compatible APIs.
• Support channels: Provide answers through a REST API, Discord, Slack or an embedded web widget, and manage knowledge sources through the API.

Ragpi runs in Docker and is open source under the MIT license. Its API and knowledge store can run on your server, with Postgres and pgvector or Redis for storage. Basic OpenTelemetry tracing helps inspect requests.

Model processing can stay local with Ollama; cloud providers receive the requests sent to them. Fetching remote sources and using community integrations requires network access. The optional OpenAI Responses API path retains conversation state in OpenAI's stored workflow and doesn't support zero-data-retention operation.