All notable changes to the MCP-Pinecone project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
- Added
pinecone-store
prompt to store documents in Pinecone - Added
pinecone-stats
tool to get stats about the Pinecone index
- Refactoring across the codebase to make it more modular and easier to extend
- Removed
embed-document
tool - Removed
chunk-document
tool
- Updated prompt to
pinecone-query
because Zed can't use tools.
- Added
chunk_enabled
argument toprocess-document
tool to enable/disable chunking. Defaults to false. - Added
list-documents
tool to list all documents in a namespace
- Added
process-document
tool to combine chunking, embedding, and upserting documents into Pinecone - Added
chunk-document
tool to explicitly chunk documents into chunks - Added
embed-document
tool to explicitly embed documents into Pinecone - Mention Pinecone api in README
- Added
langchain
dependency for chunking - Auto chunk documents by markdown headers
- Namespace support for all vector operations (search, read, upsert)
- Explicit namespace parameter in tool schemas
- Updated MCP package to latest version
- Initial public release
- Basic Pinecone integration with MCP
- Semantic search capabilities
- Document reading and writing
- Metadata support