Tool reference · Save & recall

memory_index_passages

Builds the passage index for one project in batches, which evidence-mode recall and memory_answer search over.

local stdio server writes

Caution. Computes embeddings for every record in the batch, so large batches take noticeable time.

When to use

Before using memory_recall with mode: 'evidence' or memory_answer on a project for the first time.
After importing or saving many records into a project.

Name	Type	Default	Description
`project`required	string	—	(min length 1)
`limit`	integer	`20`	(min 1, max 100)
`after_id`	integer	`0`	(min 0)

Arguments

{
  "project": "my-api",
  "limit": 50,
  "after_id": 0
}

Result shape

{
  "processed": 50,
  "next_after_id": 1790,
  "remaining": 53,
  "model": "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2"
}

Call again with after_id set to next_after_id until remaining is 0.

Values are illustrative; the keys follow the server's handler. MCP clients receive the result as JSON text content.

The description the server sends to your agent in tools/list, captured from the v14.7.0 source:

Build local passage indexes before evidence searches. Repeat using next_after_id until remaining=0.