embed-pinecone

To embed chunked content into Pinecone, run migrate-memory-to-pinecone.ts to push files using content-hash delta sync, comparing hashes and chunk IDs to update only changed records.

Content-hash delta sync compares file hashes and chunk IDs against existing Pinecone records to minimize re-embedding, refreshing only new or modified content during indexing.

Yes, you need Pinecone credentials and environment configuration to index chunked files, as the pipeline requires authenticated access to the Pinecone vector database to store embeddings.

Yes, you can scope Pinecone embeddings to specific folders using the --rel option and manifest, allowing selective embedding of targeted chunked content instead of processing everything.

The best way to refresh changed files in a Pinecone vector database is using delta-sync to compare content hashes and chunk IDs, ensuring only modified or new files are re-embedded.