Neura Market
chatgpt-retrieval-plugin logo

chatgpt-retrieval-plugin

Free

20.9k

FreeFree tier
Inputs: textOutputs: text
Type
Open Source
Company
OpenAI
Visit Website GitHub (40 stars)

About chatgpt-retrieval-plugin

The ChatGPT Retrieval Plugin is an open-source repository by OpenAI that provides a flexible backend for semantic search and retrieval of personal or organizational documents using natural language queries. It is designed to work with ChatGPT custom GPTs, function calling with the Chat Completions or Assistants APIs, and the deprecated ChatGPT plugins model. The plugin allows users to build custom retrieval systems with granular control over document text chunk length, embedding model, and vector database provider, offering a more customizable alternative to the native retrieval capabilities of ChatGPT and the Assistants API. The repository includes core logic for storing and querying document embeddings, documentation for various vector database providers, example configurations, and authentication methods.

Key Features

Semantic search and retrieval of documents using natural language queries
Integration with ChatGPT custom GPTs, Chat Completions API, and Assistants API
Support for multiple vector database providers (e.g., Pinecone, Weaviate, Qdrant)
Granular control over document text chunk length and embedding model/size
Open-source codebase with example configurations and authentication methods
Standalone retrieval backend deployable as a plugin or API endpoint

Pros & Cons

Pros
  • Open-source and free to use, with full access to the codebase
  • Flexible integration with multiple OpenAI APIs and custom GPTs
  • Granular control over retrieval parameters for optimized performance
  • Supports a wide range of vector database providers for scalability
  • Well-documented with examples and setup guides for various providers
Cons
  • Requires technical expertise to set up and deploy (e.g., Docker, vector databases)
  • Not a standalone application; needs integration with OpenAI services or custom frontends
  • Deprecated plugin model may limit future compatibility; newer APIs are recommended
  • Performance and accuracy depend on the chosen vector database and embedding model
  • Free tier of vector database providers may have usage limits that should be verified

Best For

Building custom ChatGPT GPTs that can access personal or work documentsEnabling natural language querying of organizational knowledge basesCreating a retrieval-augmented generation (RAG) pipeline for custom applicationsProviding a backend for semantic search in enterprise document management systemsDeveloping personalized assistants that retrieve information from user-provided files

FAQ

What is the ChatGPT Retrieval Plugin?
It is an open-source repository by OpenAI that provides a backend for semantic search and retrieval of documents using natural language queries, designed to work with ChatGPT custom GPTs and OpenAI APIs.
Do I need to pay to use the ChatGPT Retrieval Plugin?
The plugin code is open-source and free to use. However, deploying it may involve costs for vector database services or OpenAI API usage, which should be verified based on your chosen providers.
Can I use the Retrieval Plugin with the Assistants API?
Yes, the plugin can be used with function calling via the Chat Completions or Assistants APIs, though OpenAI notes that native retrieval from uploaded files is also available.
What vector databases are supported?
The repository supports multiple vector database providers, including Pinecone, Weaviate, Qdrant, and others. Specific setup instructions are provided in the documentation.
Is the Retrieval Plugin still actively maintained?
Based on the repository, it appears to be maintained by OpenAI, but users should check the latest commit history and release notes for current status.
Can I customize the embedding model used for retrieval?
Yes, the plugin allows granular control over the embedding model and document text chunk length, as stated in the repository documentation.