---
title: "Multi-Pass Extraction · agent retrieval pattern · Agent Fieldbook"
url: https://agentfieldbook.org/memory/retrieval/multi-pass-extraction/
description: "A full pass chunks the conversation (concurrent) for broad coverage; a detail pass runs overlapping windows to catch concrete values (names, prices, versions); results merge."
section: "Memory · Retrieval"
source: Agent Fieldbook — generated from the published page
---

# Multi-Pass Extraction

A full pass chunks the conversation (concurrent) for broad coverage; a detail pass runs overlapping windows to catch concrete values (names, prices, versions); results merge.

Verified against [blog.cloudflare.com](https://blog.cloudflare.com/introducing-agent-memory/)

- [blog.cloudflare.com](https://blog.cloudflare.com/introducing-agent-memory/)

## How it works

## At a glance

| Field | Value |
| --- | --- |
| Stage | Harvest |
| Phase | Write |
| Used by | Cloudflare Agent Memory |

## Trade-off

Multiple LLM passes per conversation, extraction cost scales with length.

## Basis

## Verified against

- [https://blog.cloudflare.com/introducing-agent-memory/](https://blog.cloudflare.com/introducing-agent-memory/)

The corpus does not rewrite vendor documentation. Where copy is quoted verbatim it is marked as such and attributed to its source.