tkmind_go/documentation/docs/mcp/memory-mcp.md

---
title: Memory Extension
description: Use Memory MCP Server as a goose Extension
---

import Tabs from '@theme/Tabs';
import TabItem from '@theme/TabItem';
import YouTubeShortEmbed from '@site/src/components/YouTubeShortEmbed';
import GooseBuiltinInstaller from '@site/src/components/GooseBuiltinInstaller';

<YouTubeShortEmbed videoUrl="https://youtube.com/embed/BZ0yrSLXQwk" />

The Memory extension turns goose into a knowledgeable assistant by allowing you to teach it personalized key information (e.g. commands, code snippets, preferences and configurations) that it can recall and apply later. Whether it’s project-specific (local) or universal (global) knowledge, goose learns and remembers what matters most to you.

This tutorial covers enabling and using the Memory MCP Server, which is a built-in goose extension.  

## Configuration

<Tabs groupId="interface">
  <TabItem value="ui" label="goose Desktop" default>
  <GooseBuiltinInstaller
    extensionName="Memory"
    description="Store and recall personalized information for consistent assistance"
  />
  </TabItem>
  <TabItem value="cli" label="goose CLI">

 
  1. Run the `configure` command:
  ```sh
  goose configure
  ```

  2. Choose to `Toggle Extensions`
  ```sh
  ┌   goose-configure 
  │
  ◇  What would you like to configure?
  │  Toggle Extensions 
  │
  ◆  Enable extensions: (use "space" to toggle and "enter" to submit)
  // highlight-start    
  │  ● memory
  // highlight-end
  |
  └  Extension settings updated successfully
  ```
  </TabItem>
</Tabs>

## Why Use Memory?  
With the Memory extension, you’re not just storing static notes, you’re teaching goose how to assist you better. Imagine telling goose:  

> _learn everything about MCP servers and save it to memory._

Later, you can ask:
> _utilizing our MCP server knowledge help me build an MCP server._ 

goose will recall everything you’ve saved as long as you instruct it to remember. This makes it easier to have consistent results when working with goose.

goose loads all saved memories at the start of a session and includes them in every prompt sent to the LLM. For large or detailed instructions, store them in files and instruct goose to reference those files:

> _Remember that if I ask for help writing JavaScript, I want you to refer to "/path/to/javascript_notes.txt" and follow the instructions in that file._


## Trigger Words and When to Use Them
goose also recognizes certain trigger words that signal when to store, retrieve, or remove memory.

| **Trigger Words**   | **When to Use** |
|---------------------|----------------|
| remember            | Store useful info for later use |
| forget           | Remove a stored memory |
| memory           | General memory-related actions |
| save             | Save a command, config, or preference |
| remove memory    | Delete specific stored data |
| clear memory     | Wipe all stored memories |
| search memory    | Find previously stored data |
| find memory      | Locate specific saved information |

## Example Usage

In this example, I’ll show you how to make goose a knowledgeable development assistant by teaching it about your project’s API standards. With the Memory extension, goose can store structured information and recall it when needed, helping you stay consistent in your work.

This means you no longer have to repeat yourself. goose will remember your project’s requirements and automatically apply them to new tasks.

:::tip  
If you frequently work with API standards or other structured knowledge, goose may automatically suggest saving them to memory for you when it detects a pattern in your workflow.  
:::

<Tabs groupId="interface">
  <TabItem value="ui" label="goose Desktop" default>
   1. Open a new session in goose Desktop
  </TabItem>
  <TabItem value="cli" label="goose CLI">

  1. Open a terminal and start a new goose session:

  ```sh
  goose session
  ```

  </TabItem>
</Tabs>

### Step 1: Teach goose Your API Standards
    #### goose Prompt #1
    ```
    goose, learn our API development standards:

    - All endpoints must use TypeScript interfaces
    - Error responses follow RFC 7807 Problem Details spec
    - Rate limiting headers are required
    - Authentication uses JWT in Authorization header
    - All endpoints need OpenAPI documentation

    Remember this so you can help me follow these practices.
    ```

    #### goose Output

    ```
    I've learned your API standards. I'll help enforce these requirements when working on API endpoints.

    Storing in:
        - Category: development_standards
        - Tags: #api #typescript #security #documentation
        - Scope: global (these are general standards)

    I can now help with:
        - API endpoint implementation
        - Error handling patterns
        - Authentication flows
        - Documentation requirements
    ```
### Step 2: Use Stored Knowledge to Create a New API Endpoint
    #### goose Prompt # 2

    ```
    Help me create a new endpoint for user preferences
    ```
    #### goose Output
    ```
    I see you need a new endpoint. Based on our API development standards that I have stored in memory, I'll ensure it meets all our requirements.

    [.. Checking stored API standards ...]

    I've included all our required elements:
    
    ✅ TypeScript interfaces
    ✅ RFC 7807 error response format
    ✅ Rate limiting
    ✅ JWT auth header
    ✅ OpenAPI documentation

    Here's your compliant endpoint:

    interface UserPreferences {
        userId: string;
        theme: 'light' | 'dark';
        notifications: boolean;
        timezone: string;
    }

    interface ErrorResponse { 
        type: string;
        title: string;
        status: number;
        detail: string;
        instance: string;
    }

    @Route('/api/users/:userId/preferences')
    class UserPreferencesController {
        @Get()
        @RateLimit(100)  // Adding required rate limiting
        @OpenAPI({ ... }) // Including mandatory OpenAPI docs
        async getUserPreferences(
            @Header('Authorization') token: string,  // JWT auth as per our standard
            @Param('userId') userId: string
        ): Promise<UserPreferences | ErrorResponse> {
            // Implementation following our error handling pattern
        }
    }

    Would you like me to implement the full endpoint logic?
    ```