> ## Documentation Index
> Fetch the complete documentation index at: https://docs.recallio.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Memory Overview

> See how Recallio captures, searches, and manages conversational memories

Recallio Memory stores conversational or assistant-derived facts so your automations can reference prior interactions. Use these pages to understand how the Memory endpoints complement the Knowledge Base.

## Core capabilities

* **Store** user-specific memories with `/api/Memory/write`, tagging and scoping them per user or project.
* **Retrieve** relevant facts in real time with `/api/Memory/recall` and refine results with thresholds and tags.
* **Summarize & explore** topics with `/api/Memory/recall-summary` and `/api/Memory/recall-topics` to power profile views.
* **Administer** the dataset with `/api/Memory/delete` and `/api/Memory/export` (Manager role required).

<Tip>
  All Memory endpoints require an API key from **Dashboard → API Keys**. Use a key with the **Manager** role for delete or export operations.
</Tip>

## Endpoint quick reference

| Use case                                | Endpoint                     | Method   | Key inputs                                                                                                                                   |
| --------------------------------------- | ---------------------------- | -------- | -------------------------------------------------------------------------------------------------------------------------------------------- |
| Add a new memory for a user or project  | `/api/Memory/write`          | `POST`   | Body fields: `userId`, `content`, `consentFlag`, optional `projectId`, `tags`, `metadata`                                                    |
| Search for the most relevant memories   | `/api/Memory/recall`         | `POST`   | Query params: `summarized`, `reRank`, `type`, `similarityThreshold`, `limit`; body: `query`, `scope`, `userId`, `projectId`, optional `tags` |
| Get a condensed view of all facts       | `/api/Memory/recall-summary` | `POST`   | Body: `scope`, `userId`, `projectId`, optional `tags`                                                                                        |
| Discover topics extracted from memories | `/api/Memory/recall-topics`  | `POST`   | Body: `userId`, optional `projectId`                                                                                                         |
| Remove memories for a scope             | `/api/Memory/delete`         | `DELETE` | Body: `scope`, optional `userId`, `projectId`                                                                                                |
| Export memories for compliance or BI    | `/api/Memory/export`         | `GET`    | Query params: `Type`, `Format`, optional `UserId`, `ProjectId`, `StartDate`, `EndDate`                                                       |

## Before you start

1. **Generate an API key** and store it securely. Keys map to a role; choose **Manager** if you need deletion or export access.
2. **Align on consent**. The `consentFlag` field must be `true` for every stored memory to meet compliance requirements.
3. **Decide on scoping**. Use `projectId` to partition memories per environment, and unique `userId` values for personalized recall.

## Where to go next

<CardGroup>
  <Card title="Memory Kick Start" icon="rocket" href="/memory-quickstart">
    Learn how to write your first memory and retrieve it with `/api/Memory/recall`.
  </Card>

  <Card title="Memory Deep Dive" icon="code" href="/memory-deep-dive">
    Explore payload schemas, query tuning, and admin workflows for Memory.
  </Card>
</CardGroup>

Need the full schema? Refer to the [Recallio Swagger reference](https://app.recallio.ai/swagger/index.html).