Skip to content
forked from ragpi/ragpi

πŸ€– An API-first AI assistant using Agentic RAG to search and answer questions from technical sources

License

Notifications You must be signed in to change notification settings

bhardwajRahul/ragpi

Β 
Β 

Repository files navigation

Ragpi

Ragpi is an open-source AI assistant that answers questions using your documentation, GitHub issues, and READMEs. It combines LLMs with intelligent search to provide relevant, documentation-backed answers through a simple API. It supports multiple providers like OpenAI, Ollama, and Deepseek, and has built-in integrations with Discord and Slack. A web widget integration is also available to embed the assistant in your website.

Documentation | API Reference

Key Features

  • πŸ“š Builds knowledge bases from docs, GitHub issues and READMEs
  • πŸ€– Agentic RAG system for dynamic document retrieval
  • πŸ”Œ Supports OpenAI, Ollama, Deepseek & OpenAI-Compatible models
  • πŸ’¬ Discord and slack integrations for community support
  • πŸš€ API-first design with Docker deployment

Example Workflow

Here's a simple workflow to get started with Ragpi once it's deployed:

1. Set up a Source with a Connector

  • Use the /sources endpoint to configure a source with your chosen connector.
  • Each connector type has its own configuration parameters.

Example payload using the Sitemap connector:

{
  "name": "example-docs",
  "description": "Documentation for example project. It contains information about configuration, usage, and deployment.",
  "connector": {
    "type": "sitemap",
    "sitemap_url": "https://docs.example.com/sitemap.xml"
  }
}

2. Monitor Source Synchronization

  • After adding a source, documents will be synced automatically. You can monitor the sync process through the /tasks endpoint.

3. Chat with the AI Assistant

  • Use the /chat endpoint to query the AI assistant using the configured sources:

    {
      "sources": ["example-docs"],
      "messages": [
        { "role": "user", "content": "How do I deploy the example project?" }
      ]
    }
  • You can also interact with the AI assistant through the Discord or Slack integration, or by embedding the Web Widget in your website.

Connectors

Ragpi supports the following connectors for building knowledge bases:

  • Documentation Website (Sitemap)
  • GitHub Issues
  • GitHub README Files
  • GitHub PDF Files
  • REST API Responses

Explore connectors β†’

Providers

Ragpi supports the following LLM providers for generating responses and embeddings:

  • OpenAI (default)
  • Ollama
  • Deepseek
  • OpenAI-compatible APIs

Configure providers β†’

Integrations

Ragpi supports the following integrations for interacting with the AI assistant:

Contributing

Contributions to Ragpi are welcome! Please check out the contributing guidelines for more information.

About

πŸ€– An API-first AI assistant using Agentic RAG to search and answer questions from technical sources

Resources

License

Contributing

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 99.7%
  • Dockerfile 0.3%