●Change8

v1.0.0

Breaking Changes
📦 privategptView on GitHub →
⚠ 2 breaking✨ 10 features

Summary

PrivateGPT 1.0 merges two years of development, transforming the project from a semantic search pipeline into a full application backend that connects to external inference servers for production AI applications.

⚠️ Breaking Changes

  • The original PrivateGPT API is not forward-compatible with 1.0. Consult the migration guide in the documentation for necessary changes.
  • PrivateGPT 1.0 no longer runs models internally; it requires connection to an external OpenAI-compatible inference server (e.g., Ollama, llama.cpp, vLLM).

Migration Steps

  1. Review the migration guide in the documentation to update existing code from the previous PrivateGPT API to the new 1.0 API.
  2. Configure PrivateGPT 1.0 to connect to an external OpenAI-compatible inference server, as model execution is no longer internal.

✨ New Features

  • Standard messages API introduced.
  • Added files and artifact ingestion capabilities.
  • Implemented retrieval with citations and agentic RAG.
  • Built-in tools mirroring those offered by the Claude API.
  • Support for custom tools.
  • Introduction of MCP connectors.
  • Structured access to databases and CSVs.
  • Added web search and extraction functionality.
  • Support for code execution.
  • Included token counting, embeddings, and orchestration features.