PlugnPlai
Simplifies the integration of AI plugins and external tools into open-source large language models (LLMs).
About
PlugnPlai is an open-source Python library designed to streamline the process of connecting external AI plugins with large language models. It offers comprehensive utilities for discovering, loading, and managing plugins from a central directory, including fetching their manifests and OpenAPI specifications. The library facilitates dynamic interaction by enabling parsing of LLM responses for API calls, executing those calls, and even providing a decorator for seamless plugin application within LLM functions. It also supports plugin retrieval via a vector database, making it easier for LLMs to find and utilize relevant tools.
Key Features
- Activate and deactivate plugins, generating contextual prompts for LLMs.
- Execute API calls to active plugins, supporting various authentication methods.
- Retrieve relevant plugins using a vector database based on user messages.
- Load plugin manifests and OpenAPI specifications for integration.
- Parse LLM responses to identify and extract API calls (plugin name, operation ID, parameters).
- Discover and retrieve AI plugins from a directory, with filtering options.
- 230 GitHub stars
Use Cases
- Build custom LLM agents capable of utilizing a diverse set of tools and APIs.
- Automate the process of integrating and managing AI functionalities within LLM applications.
- Enable open-source LLMs to interact with external services and data via plugins.