Ingest full documents

[moltar]

Motivation

I have a large local text file that I want to ingest into the index.

It'd be really nice if MCP supported this use case, without having to write any code to create chunks and then feed them one by one to upsert.

Suggested solution

N/A

Alternatives

Writing own code.

Additional context

N/A

[avi1mizrahi]

Are you familiar with Pinecone Assistant?
It does exactly that. For query, you have two options, chat-based API and context API which retrieves chunks based on your query that you can feed into your LLM, or use it directly via one of our MCP servers