---
title: "Extract Memories From Conversation"
method: POST
path: "/api/v2/agents/{agent_id}/remember/extract"
tags: ["Sessions & Agents", "Memory Operations"]
---

# Extract Memories From Conversation

`POST /api/v2/agents/{agent_id}/remember/extract`

Extract typed memory candidates from chat-style conversation turns.

Requires:
- X-Session-Token: {session_token}

When dry_run is true, candidates are returned without writing. Otherwise the
candidates are persisted through the same batch memory path used by
/batch-remember.

## Path parameters

- `agent_id` string, required

## Headers

- `x-session-token` string, nullable
- `X-Api-Key` string, nullable

## Cookies

- `memanto_session_token` string, nullable

## Request body

- ExtractMemoriesRequest — Request body for extracting typed memories from conversation turns.
  - `messages` ConversationMessage[], required — Conversation turns to extract durable memories from
    - `role` string, required
    - `content` string, required
  - `dry_run` boolean — When true, return candidates without storing them
  - `max_memories` integer — Maximum candidate memories to extract
  - `ai_model` string, nullable — Optional model override for extraction

## Response `200`

Successful Response

- unknown

## Other responses

- `422` — Validation Error

---

[API](https://skmtc.net/moorcheh-ai/apis/memanto-memory-that-ai-agents-love.md) · [All operations](https://skmtc.net/moorcheh-ai/apis/memanto-memory-that-ai-agents-love/llms.txt) · [OpenAPI document](https://skmtc-service-staging.skmtc.workers.dev/v1/apis/moorcheh-ai/memanto-memory-that-ai-agents-love/revisions/f8fc41287ced/schema)