### Clone Repository and Install Dependencies

Source: https://github.com/aristoapp/openclaw-membase/blob/main/CONTRIBUTING.md

Use these commands to clone the repository and install project dependencies using Bun.

```bash
git clone https://github.com/aristoapp/openclaw-membase.git
cd openclaw-membase
bun install
```

--------------------------------

### Install Membase Plugin

Source: https://github.com/aristoapp/openclaw-membase/blob/main/README.md

Install the Membase plugin for OpenClaw using the provided command. Restart OpenClaw after installation.

```bash
openclaw plugins install @membase/openclaw-membase
```

--------------------------------

### Clone and Install OpenClaw Membase

Source: https://github.com/aristoapp/openclaw-membase/blob/main/README.md

Set up the development environment by cloning the repository, installing dependencies, and running type checks, linting, and build processes.

```bash
git clone https://github.com/aristoapp/openclaw-membase.git
cd openclaw-membase
bun install
bun run check-types
bun run lint
bun run build
```

--------------------------------

### Install Membase Plugin for OpenClaw

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Install the Membase plugin using the OpenClaw plugin manager. Remember to restart OpenClaw and authenticate afterwards.

```bash
openclaw plugins install @membase/openclaw-membase

# Restart OpenClaw after installation
# Then authenticate with OAuth
openclaw membase login
```

--------------------------------

### Setup Membase Login

Source: https://github.com/aristoapp/openclaw-membase/blob/main/README.md

Log in to Membase using OAuth authentication. This command opens a browser for authentication, and tokens are saved automatically.

```bash
openclaw membase login
```

--------------------------------

### Get User Profile

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Retrieve the user's profile settings from Membase, including display name, role, interests, and instructions.

```typescript
const profile = await client.getProfile();
console.log(`User: ${profile.display_name}`);
console.log(`Role: ${profile.role}`);
console.log(`Interests: ${profile.interests}`);
console.log(`Instructions: ${profile.instructions}`);
```

--------------------------------

### Get User Profile Memory

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Fetch the user's profile memory node from the knowledge graph. Returns the memory node if found, otherwise null.

```typescript
const profileMemory = await client.getUserProfileMemory();
if (profileMemory) {
  console.log(`Profile UUID: ${profileMemory.episode.uuid}`);
  console.log(`Profile: ${profileMemory.episode.summary}`);
}
```

--------------------------------

### Retrieve User Profile and Related Memories

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Fetches the user's profile information and associated memories. Ideal for initializing session context at the start of a new session.

```typescript
// Get user profile with related memories
const profile = await membase_profile({});

// Response includes:
// - User profile (display_name, role, interests, instructions)
// - Related memories from knowledge graph
// - User profile memory node
```

--------------------------------

### membase_profile

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Retrieve the user's profile and related memories for session context. Best used at the start of new sessions for an overview of the user.

```APIDOC
## membase_profile

### Description
Retrieve the user's profile and related memories for session context. Best used at the start of new sessions for an overview of the user.

### Method
GET

### Endpoint
/membase/profile

### Parameters
No parameters are required for this endpoint.

### Request Example
```json
{}
```

### Response
#### Success Response (200)
- **profile** (object) - Contains user profile information.
  - **display_name** (string) - The user's display name.
  - **role** (string) - The user's role.
  - **interests** (array of strings) - User's interests.
  - **instructions** (string) - Specific instructions for the user.
- **related_memories** (array) - Memories related to the user's profile.
- **profile_memory_node** (object) - The knowledge graph node representing the user's profile.

#### Response Example
```json
{
  "profile": {
    "display_name": "Jane Doe",
    "role": "Software Engineer",
    "interests": ["AI", "Machine Learning"],
    "instructions": "Focus on efficiency and scalability."
  },
  "related_memories": [
    {
      "id": "uuid-related-1",
      "content": "Worked on project X."
    }
  ],
  "profile_memory_node": {
    "id": "uuid-profile-node"
  }
}
```
```

--------------------------------

### Initialize MembaseClient

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Create a new MembaseClient instance for direct API interaction. Requires API URL, authentication tokens, and optional configuration like debug mode and token refresh callbacks.

```typescript
import { MembaseClient } from "./client";

const client = new MembaseClient(
  "https://api.membase.so",  // API URL
  {
    accessToken: "your-access-token",
    refreshToken: "your-refresh-token",
    clientId: "your-client-id"
  },
  {
    debug: true,
    timeoutMs: 15000,
    onTokenRefresh: (tokens) => {
      console.log("Tokens refreshed:", tokens.accessToken.slice(0, 10) + "...");
      // Persist new tokens
    }
  }
);

// Check authentication status
if (client.isAuthenticated()) {
  console.log("Client ready for API calls");
}
```

--------------------------------

### MembaseClient Constructor

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Initializes a new MembaseClient instance for interacting with the Membase backend. Requires API URL and authentication tokens.

```APIDOC
## MembaseClient Constructor

### Description
Create a new MembaseClient instance for direct API interaction with the Membase backend.

### Request Example
```typescript
import { MembaseClient } from "./client";

const client = new MembaseClient(
  "https://api.membase.so",  // API URL
  {
    accessToken: "your-access-token",
    refreshToken: "your-refresh-token",
    clientId: "your-client-id"
  },
  {
    debug: true,
    timeoutMs: 15000,
    onTokenRefresh: (tokens) => {
      console.log("Tokens refreshed:", tokens.accessToken.slice(0, 10) + "...");
      // Persist new tokens
    }
  }
);

// Check authentication status
if (client.isAuthenticated()) {
  console.log("Client ready for API calls");
}
```
```

--------------------------------

### Add Wiki Document via CLI

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Create new wiki documents for storing factual knowledge. Documents can include content, be assigned to specific collections, and optionally be summarized.

```bash
# Add a basic wiki document
openclaw membase wiki-add "Project Setup Guide" --content "# Setup\n\nInstall dependencies with npm install"

# Add to specific collection
openclaw membase wiki-add "API Reference" --content "## Endpoints\n\n- GET /users" --collection-id api-docs

# Add with automatic summarization
openclaw membase wiki-add "Meeting Notes" --content "Long meeting content..." --summarize
```

--------------------------------

### MembaseClient.createWikiDocument

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Creates a new wiki document, with options for specifying a collection and enabling summarization.

```APIDOC
## MembaseClient.createWikiDocument

### Description
Create a new wiki document with optional collection and summarization.

### Method
`client.createWikiDocument(title: string, content: string, collection?: string, summarize?: boolean): Promise<WikiDocument>

### Parameters
#### Request Body
- **title** (string) - Required - The title of the wiki document.
- **content** (string) - Required - The content of the wiki document.
- **collection** (string) - Optional - The collection to place the document in.
- **summarize** (boolean) - Optional - Whether to automatically summarize the content.

### Response
#### Success Response (200)
- **title** (string) - The title of the created document.
- **id** (string) - The unique identifier of the created document.

### Request Example
```typescript
const doc = await client.createWikiDocument(
  "Deployment Checklist",           // title
  "# Pre-deployment\n- Run tests\n- Review PR\n\n# Deployment\n- Deploy to staging\n- Smoke test",  // content
  "devops",                         // collection (optional)
  true                              // summarize (optional)
);
console.log(`Created: ${doc.title} (${doc.id})`);
```
```

--------------------------------

### Search Membase Wiki via CLI

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Search the Membase knowledge wiki for factual references. Results can be limited and filtered by collection.

```bash
# Search wiki documents
openclaw membase wiki-search "API authentication"

# Limit results
openclaw membase wiki-search "deployment" --limit 5

# Filter by collection
openclaw membase wiki-search "configuration" --collection-id work-docs
```

--------------------------------

### Create Wiki Document

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Create a new wiki document with optional collection and summarization. Logs the title and ID of the created document.

```typescript
const doc = await client.createWikiDocument(
  "Deployment Checklist",           // title
  "# Pre-deployment\n- Run tests\n- Review PR\n\n# Deployment\n- Deploy to staging\n- Smoke test",  // content
  "devops",                         // collection (optional)
  true                              // summarize (optional)
);
console.log(`Created: ${doc.title} (${doc.id})`);
```

--------------------------------

### Run Type Checks and Linting

Source: https://github.com/aristoapp/openclaw-membase/blob/main/CONTRIBUTING.md

Execute these commands to ensure code quality and type correctness before committing changes.

```bash
bun run check-types
```

```bash
bun run lint
```

--------------------------------

### Configure OpenClaw Membase Plugin

Source: https://github.com/aristoapp/openclaw-membase/blob/main/README.md

Manually configure the plugin by adding 'openclaw-membase' to 'tools.alsoAllow' and specifying plugin-specific settings. This ensures the AI can access memory tools and customizes recall behavior.

```json
{
  "tools": {
    "profile": "coding",
    "alsoAllow": ["openclaw-membase"]
  },
  "plugins": {
    "entries": {
      "openclaw-membase": {
        "enabled": true,
        "config": {
          "autoRecall": true,
          "autoWikiRecall": false,
          "autoCapture": true,
          "maxRecallChars": 4000
        }
      }
    }
  }
}
```

--------------------------------

### MembaseClient.ingest

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Stores content to memory, with optional parameters for display summary and project categorization.

```APIDOC
## MembaseClient.ingest

### Description
Store content to memory with optional display summary and project categorization.

### Method
`client.ingest(content: string, options?: { displaySummary?: string; project?: string }): Promise<{ status: string }>

### Parameters
#### Request Body
- **content** (string) - Required - The content to ingest.
- **options** (object) - Optional - Additional options.
  - **displaySummary** (string) - Optional - A summary to display for the ingested content.
  - **project** (string) - Optional - The project to categorize the content under.

### Request Example
```typescript
const result = await client.ingest(
  "User completed the React certification course on March 1st, 2026.",
  {
    displaySummary: "React certification completed",
    project: "learning"
  }
);
console.log(`Ingestion status: ${result.status}`);
```
```

--------------------------------

### Authenticate with Membase CLI

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Authenticate with Membase using the CLI. The standard command initiates an OAuth 2.0 PKCE flow opening a browser. Custom API URLs and callback ports can be specified.

```bash
# Standard login (opens browser automatically)
openclaw membase login

# Custom API URL for self-hosted instances
openclaw membase login --api-url https://custom.membase.example.com

# Specify callback port if default (8765) is in use
openclaw membase login --port 9000
```

--------------------------------

### Search Membase Memories via CLI

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Perform semantic searches on stored memories using the CLI. Results can be limited and filtered by source.

```bash
# Basic semantic search
openclaw membase search "project deadlines"

# Search with result limit
openclaw membase search "meetings" --limit 20

# Filter by specific sources
openclaw membase search "team updates" --sources slack,gmail
```

--------------------------------

### Update Wiki Document via CLI

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Update existing wiki documents by modifying their title, content, or collection using the document ID.

```bash
# Update document title
openclaw membase wiki-update abc123 --title "Updated Project Guide"

# Update document content
openclaw membase wiki-update abc123 --content "# New Content\n\nUpdated information"

# Move to different collection
openclaw membase wiki-update abc123 --collection api-reference
```

--------------------------------

### Membase CLI Commands

Source: https://github.com/aristoapp/openclaw-membase/blob/main/README.md

Common CLI commands for interacting with Membase, including login, logout, searching memories and wiki documents, and managing wiki content.

```bash
openclaw membase login              # OAuth login (PKCE) — opens browser
```

```bash
openclaw membase logout             # Remove stored tokens
```

```bash
openclaw membase search <query>     # Search memories
```

```bash
openclaw membase search <query> -s slack,gmail  # Filter by source
```

```bash
openclaw membase wiki-search <query>            # Search wiki documents
```

```bash
openclaw membase wiki-add "<title>" --content "<markdown>"   # Add wiki doc
```

```bash
openclaw membase wiki-update <docId> --title "<new title>"   # Update wiki doc
```

```bash
openclaw membase wiki-delete <docId>            # Delete wiki doc
```

```bash
openclaw membase status             # Check API connectivity
```

--------------------------------

### Check Membase API Status via CLI

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Verify the connectivity status of the Membase API using the status command.

```bash
# Verify API connection
openclaw membase status
```

--------------------------------

### Configure Membase Plugin in OpenClaw

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Configure the Membase plugin settings either through OpenClaw's plugin settings interface or by directly editing the `~/.openclaw/openclaw.json` configuration file. This JSON defines plugin enablement, API URL, token file location, and various recall/capture settings.

```json
{
  "tools": {
    "profile": "coding",
    "alsoAllow": ["openclaw-membase"]
  },
  "plugins": {
    "entries": {
      "openclaw-membase": {
        "enabled": true,
        "config": {
          "apiUrl": "https://api.membase.so",
          "tokenFile": "~/.openclaw/credentials/openclaw-membase.json",
          "autoRecall": true,
          "autoWikiRecall": false,
          "autoCapture": true,
          "maxRecallChars": 4000,
          "debug": false
        }
      }
    }
  }
}
```

--------------------------------

### Membase AI Tools

Source: https://github.com/aristoapp/openclaw-membase/blob/main/README.md

Overview of AI tools provided by the Membase plugin for interacting with memory and wiki documents. These tools can be used autonomously by the agent.

```markdown
| Tool | Description |
| --- | --- |
| `membase_search` | Search memories by semantic similarity. Supports date filtering (`date_from`, `date_to`, `timezone`) and source filtering (`sources` — e.g. `['slack', 'gmail']`). Returns episode bundles with related facts and relevance scores. |
| `membase_store` | Save important information to long-term memory. Proactively stores preferences, goals, and context. |
| `membase_forget` | Delete a memory. Shows matches first, then deletes after user confirmation (two-step). |
| `membase_profile` | Retrieve user profile and related memories for session context. |
| `membase_search_wiki` | Search wiki documents for factual references and stable knowledge. |
| `membase_add_wiki` | Create a wiki document from markdown content. |
| `membase_update_wiki` | Update title/content/collection of an existing wiki document. |
| `membase_delete_wiki` | Delete a wiki document with confirmation flow. |
```

--------------------------------

### Handle MembaseApiError

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Demonstrates how to catch and handle MembaseApiError, checking the status code for specific actions like re-authentication or rate limit handling.

```typescript
import { MembaseApiError } from "./types";

try {
  const result = await client.search("test query");
} catch (error) {
  if (error instanceof MembaseApiError) {
    console.error(`API Error: ${error.message}`);
    console.error(`Status: ${error.status}`);
    console.error(`Body: ${error.body}`);

    if (error.status === 401) {
      console.log("Re-authentication required. Run 'openclaw membase login'");
    } else if (error.status === 429) {
      console.log("Rate limited. Please wait before retrying.");
    }
  } else {
    console.error("Unexpected error:", error);
  }
}
```

--------------------------------

### Hooks System

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Documentation for the Hooks System, including Auto-Recall and Auto-Capture hooks.

```APIDOC
## Hooks System

### Auto-Recall Hook

The recall hook automatically injects relevant memories before each AI turn, searching both episodic memories and wiki documents based on the user's message.

**Configuration Options:**
- `autoRecall` (boolean): Enable memory recall.
- `autoWikiRecall` (boolean): Enable wiki recall.
- `maxRecallChars` (number): Context budget (500-16000).

**Functionality:**
1. Extracts the user's last message.
2. Skips casual chat and short messages.
3. Searches memories (limit: 10) and wiki (limit: 5) in parallel.
4. Formats results within character budget.
5. Injects as prependContext to the AI turn.

### Auto-Capture Hook

The capture hook automatically buffers conversation messages and flushes them to Membase for entity extraction after periods of silence or buffer limits.

**Behavior:**
- Buffers user messages from each conversation.
- Filters out heartbeat/operational messages.
- Flushes after 5 minutes of silence.
- Flushes when buffer reaches 20 messages.
- Minimum 2 messages required to flush.

**Manual Flush:**
`flushAllBuffers(client, logger)` can be used to manually trigger a flush, typically on shutdown.
```

--------------------------------

### MembaseClient.getProfile

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Retrieves the user's profile settings from Membase.

```APIDOC
## MembaseClient.getProfile

### Description
Retrieve the user's profile settings from Membase.

### Method
`client.getProfile(): Promise<UserProfile>

### Response
#### Success Response (200)
- **display_name** (string) - The user's display name.
- **role** (string) - The user's role.
- **interests** (string[]) - A list of the user's interests.
- **instructions** (string) - Special instructions for the user.

### Request Example
```typescript
const profile = await client.getProfile();
console.log(`User: ${profile.display_name}`);
console.log(`Role: ${profile.role}`);
console.log(`Interests: ${profile.interests}`);
console.log(`Instructions: ${profile.instructions}`);
```
```

--------------------------------

### membase_search_wiki

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Search the user's knowledge wiki using hybrid semantic and keyword matching. Use for factual knowledge, references, and stable documentation rather than personal context.

```APIDOC
## membase_search_wiki

### Description
Search the user's knowledge wiki using hybrid semantic and keyword matching. Use for factual knowledge, references, and stable documentation rather than personal context.

### Method
POST

### Endpoint
/membase/search_wiki

### Parameters
#### Request Body
- **query** (string) - Required - The search query string. An empty query fetches recent documents.
- **collection** (string) - Optional - Filter search results by a specific wiki collection.
- **limit** (integer) - Optional - The maximum number of results to return.

### Request Example
```json
{
  "query": "authentication flow",
  "limit": 10
}
```

### Response
#### Success Response (200)
- **results** (array) - An array of wiki documents matching the query.
  - **title** (string) - The title of the wiki document.
  - **content** (string) - The content of the wiki document.
  - **collection** (string) - The collection the document belongs to.
  - **timestamp** (string) - The timestamp of the document.

#### Response Example
```json
{
  "results": [
    {
      "title": "API Authentication Flow",
      "content": "# Authentication Flow\n\nDetails about the authentication process...",
      "collection": "api-reference",
      "timestamp": "2023-10-26T12:00:00Z"
    }
  ]
}
```
```

--------------------------------

### membase_add_wiki

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Add documents to the user's wiki knowledge base for storing factual documents and references. Supports markdown content with wikilinks.

```APIDOC
## membase_add_wiki

### Description
Add documents to the user's wiki knowledge base for storing factual documents and references. Supports markdown content with wikilinks.

### Method
POST

### Endpoint
/membase/add_wiki

### Parameters
#### Request Body
- **title** (string) - Required - The title of the wiki document.
- **content** (string) - Required - The markdown content of the wiki document.
- **collection** (string) - Optional - The collection to categorize this document under.
- **summarize** (boolean) - Optional - If true, the system will attempt to auto-generate a summary.

### Request Example
```json
{
  "title": "API Authentication Guide",
  "content": "# Authentication\n\nUse Bearer tokens for API access.\n\n## Getting a Token\n\nCall POST /auth/login with credentials.",
  "summarize": true
}
```

### Response
#### Success Response (200)
- **id** (string) - The unique identifier of the newly created wiki document.
- **title** (string) - The title of the created document.
- **collection** (string) - The collection the document was added to.

#### Response Example
```json
{
  "id": "uuid-wiki-doc-123",
  "title": "API Authentication Guide",
  "collection": "documentation"
}
```
```

--------------------------------

### MembaseClient.searchWiki

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Searches wiki documents, with an option to filter by collection.

```APIDOC
## MembaseClient.searchWiki

### Description
Search wiki documents with optional collection filtering.

### Method
`client.searchWiki(query: string, limit?: number, collection?: string): Promise<WikiSearchResult>

### Parameters
#### Query Parameters
- **query** (string) - Required - The search query.
- **limit** (number) - Optional - The maximum number of results to return.
- **collection** (string) - Optional - The collection to search within.

### Response
#### Success Response (200)
- **documents** (array) - A list of matching wiki documents.
  - **title** (string) - The title of the document.
  - **id** (string) - The unique identifier of the document.
  - **collection_name** (string) - The name of the collection the document belongs to.
  - **similarity** (number) - The similarity score of the document to the query.

### Request Example
```typescript
const wikiResult = await client.searchWiki(
  "deployment guide",  // query
  10,                  // limit
  "devops"            // collection
);

for (const doc of wikiResult.documents) {
  console.log(`${doc.title} (${doc.id})`);
  console.log(`  Collection: ${doc.collection_name}`);
  console.log(`  Similarity: ${doc.similarity}`);
}
```
```

--------------------------------

### MembaseClient.search

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Searches memories using semantic similarity, with options for date and source filtering.

```APIDOC
## MembaseClient.search

### Description
Search memories with semantic similarity, date filtering, and source filtering.

### Method
`client.search(query: string, limit?: number, offset?: number, dateFrom?: string, dateTo?: string, timezone?: string, sources?: string[], project?: string): Promise<SearchBundle[]>

### Parameters
#### Query Parameters
- **query** (string) - Required - The search query.
- **limit** (number) - Optional - The maximum number of results to return.
- **offset** (number) - Optional - The number of results to skip.
- **dateFrom** (string) - Optional - Filter results from this date (ISO 8601 format).
- **dateTo** (string) - Optional - Filter results up to this date (ISO 8601 format).
- **timezone** (string) - Optional - The timezone for date filtering.
- **sources** (string[]) - Optional - Filter results by source (e.g., ["slack", "gmail"]).
- **project** (string) - Optional - Filter results by project.

### Request Example
```typescript
const bundles = await client.search(
  "meeting notes",     // query
  20,                  // limit
  0,                   // offset
  "2026-01-01",       // dateFrom (ISO 8601)
  "2026-03-31",       // dateTo (ISO 8601)
  "America/New_York", // timezone
  ["slack", "gmail"], // sources filter
  "work-project"      // project filter
);

// Process results
for (const bundle of bundles) {
  console.log(`Memory: ${bundle.episode.name}`);
  console.log(`  Summary: ${bundle.episode.summary}`);
  console.log(`  Score: ${bundle.relevance_score}`);
  for (const edge of bundle.edges) {
    console.log(`  Fact: ${edge.fact}`);
  }
}
```
```

--------------------------------

### MembaseClient.getUserProfileMemory

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Fetches the user's profile memory node from the knowledge graph.

```APIDOC
## MembaseClient.getUserProfileMemory

### Description
Fetch the user's profile memory node from the knowledge graph.

### Method
`client.getUserProfileMemory(): Promise<SearchBundle | null>

### Response
#### Success Response (200)
- **episode** (object) - The memory episode details.
  - **uuid** (string) - The unique identifier for the memory episode.
  - **summary** (string) - A summary of the memory episode.

### Request Example
```typescript
const profileMemory = await client.getUserProfileMemory();
if (profileMemory) {
  console.log(`Profile UUID: ${profileMemory.episode.uuid}`);
  console.log(`Profile: ${profileMemory.episode.summary}`);
}
```
```

--------------------------------

### Add Documents to Wiki Knowledge Base

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Stores factual documents and references in the user's wiki. Supports markdown content, wikilinks, collection categorization, and automatic summarization.

```typescript
// Create a basic wiki document
const newDoc = await membase_add_wiki({
  title: "API Authentication Guide",
  content: "# Authentication\n\nUse Bearer tokens for API access.\n\n## Getting a Token\n\nCall POST /auth/login with credentials."
});
```

```typescript
// Create with collection categorization
const categorizedDoc = await membase_add_wiki({
  title: "Database Schema",
  content: "# Users Table\n\n| Column | Type |\n|--------|------|\n| id | UUID |\n| email | VARCHAR |",
  collection: "technical-docs"
});
```

```typescript
// Create with automatic summarization
const summarizedDoc = await membase_add_wiki({
  title: "Meeting Notes 2024-03-15",
  content: "# Product Planning Meeting\n\nLong detailed meeting notes...",
  summarize: true
});
```

--------------------------------

### Delete Wiki Document via CLI

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Delete wiki documents from Membase by providing their unique document ID.

```bash
# Delete a specific wiki document
openclaw membase wiki-delete abc123
```

--------------------------------

### Perform Semantic Search on Memories

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Use for general semantic search. Supports filtering by date range, source, and project, as well as pagination.

```typescript
const result = await membase_search({
  query: "user preferences for dark mode",
  limit: 20
});
```

```typescript
const scheduleResult = await membase_search({
  query: "",  // Empty query for broad date retrieval
  date_from: "2026-02-02",
  date_to: "2026-02-19T23:59:59",
  timezone: "Asia/Seoul",
  limit: 30
});
```

```typescript
const slackMemories = await membase_search({
  query: "team meeting",
  sources: ["slack", "gmail"],
  limit: 10
});
```

```typescript
const projectMemories = await membase_search({
  query: "deployment steps",
  project: "backend-api",
  limit: 15
});
```

```typescript
const nextPage = await membase_search({
  query: "coding patterns",
  limit: 10,
  offset: 10
});
```

--------------------------------

### Search User Knowledge Wiki

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Performs hybrid semantic and keyword matching searches on the user's wiki. Suitable for factual knowledge and documentation, not personal context.

```typescript
// Search wiki by topic
const wikiResult = await membase_search_wiki({
  query: "authentication flow",
  limit: 10
});
```

```typescript
// Search within specific collection
const apiDocs = await membase_search_wiki({
  query: "REST endpoints",
  collection: "api-reference",
  limit: 5
});
```

```typescript
// Fetch recent wiki documents
const recentDocs = await membase_search_wiki({
  query: "",  // Empty query fetches recent
  limit: 10
});
```

--------------------------------

### Error Handling

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Details on the custom MembaseApiError class for handling API-related errors.

```APIDOC
## Error Handling

### MembaseApiError

Custom error class for API-related errors with status code and response body information.

**Properties:**
- `message` (string): The error message.
- `status` (number): The HTTP status code of the error response.
- `body` (any): The response body from the API.

**Common Status Codes:**
- `401`: Re-authentication required. Run 'openclaw membase login'.
- `429`: Rate limited. Please wait before retrying.
```

--------------------------------

### MembasePluginConfig Type Definition

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Defines the configuration options for the Membase plugin, including API URL, authentication details, and settings for recall and capture hooks.

```typescript
interface MembasePluginConfig {
  apiUrl: string;        // Membase API URL
  clientId: string;      // OAuth client ID
  tokenFile: string;     // Path to token cache file
  accessToken: string;   // Current access token
  refreshToken: string;  // Current refresh token
  autoRecall: boolean;   // Enable auto-recall hook
  autoWikiRecall: boolean; // Enable wiki recall
  autoCapture: boolean;  // Enable auto-capture hook
  maxRecallChars: number; // Context budget (500-16000)
  debug: boolean;        // Enable debug logging
}
```

--------------------------------

### Update Wiki Document

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Update an existing wiki document's title, content, or collection. Use membase_search_wiki first to find the document ID.

```typescript
// Update document title
const updated = await membase_update_wiki({
  doc_id: "doc-abc123",
  title: "Updated API Guide v2"
});
```

```typescript
// Update document content
const contentUpdate = await membase_update_wiki({
  doc_id: "doc-abc123",
  content: "# Revised Content\n\nNew documentation..."
});
```

```typescript
// Move to different collection
const moveDoc = await membase_update_wiki({
  doc_id: "doc-abc123",
  collection: "archived-docs"
});
```

```typescript
// Update multiple fields
const fullUpdate = await membase_update_wiki({
  doc_id: "doc-abc123",
  title: "Complete API Reference",
  content: "# Full Documentation\n\n...",
  collection: "api-docs"
});
```

--------------------------------

### Type Definitions

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Defines the primary data structures used by the Membase API.

```APIDOC
## Type Definitions

### EpisodeBundle

The primary response type for memory searches containing an episode with related knowledge graph data.

**Properties:**
- `episode` (object): Details of the episode.
  - `uuid` (string): Unique identifier for the episode.
  - `name` (string): Name of the episode.
  - `labels` (string[]): Optional labels associated with the episode.
  - `summary` (string | null): Summary of the episode.
  - `source` (string | null): Source of the episode data.
  - `created_at` (string | null): Timestamp of creation.
  - `valid_at` (string | null): Timestamp of validity.
  - `attributes` (Record<string, unknown>): Optional attributes.
- `relevance_score` (number | null): Relevance score of the episode.
- `nodes` (NodeResponse[]): Related nodes in the knowledge graph.
- `edges` (object[]): Related edges in the knowledge graph.
  - `uuid` (string): Unique identifier for the edge.
  - `name` (string): Name of the edge.
  - `fact` (string | null): Factual information associated with the edge.
  - `source_node_uuid` (string): UUID of the source node.
  - `target_node_uuid` (string): UUID of the target node.
  - `created_at` (string | null): Timestamp of creation.

### WikiSearchDocument

Response type for wiki document searches with similarity scoring.

**Properties:**
- `id` (string): Unique identifier for the wiki document.
- `title` (string): Title of the wiki document.
- `content` (string): Content of the wiki document.
- `metadata` (Record<string, unknown>): Metadata associated with the document.
- `source` (string): Source of the document.
- `collection_id` (string | null): ID of the collection the document belongs to.
- `collection_name` (string | null): Name of the collection.
- `similarity` (number | null): Similarity score.
- `created_at` (string | null): Timestamp of creation.
- `updated_at` (string | null): Timestamp of last update.

### MembasePluginConfig

Configuration options for the Membase plugin.

**Properties:**
- `apiUrl` (string): Membase API URL.
- `clientId` (string): OAuth client ID.
- `tokenFile` (string): Path to token cache file.
- `accessToken` (string): Current access token.
- `refreshToken` (string): Current refresh token.
- `autoRecall` (boolean): Enable auto-recall hook.
- `autoWikiRecall` (boolean): Enable wiki recall.
- `autoCapture` (boolean): Enable auto-capture hook.
- `maxRecallChars` (number): Context budget (500-16000).
- `debug` (boolean): Enable debug logging.
```

--------------------------------

### membase_store

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Store important information to long-term memory that persists across sessions. Use for explicit save requests and proactive storage of durable user context like preferences, goals, and ongoing projects.

```APIDOC
## membase_store

### Description
Store important information to long-term memory that persists across sessions. Use for explicit save requests and proactive storage of durable user context like preferences, goals, and ongoing projects.

### Method
POST

### Endpoint
/membase/store

### Parameters
#### Request Body
- **content** (string) - Required - The main content of the memory to store.
- **display_summary** (string) - Optional - A brief summary or title for the memory.
- **project** (string) - Optional - The project to associate this memory with.

### Request Example
```json
{
  "content": "User prefers TypeScript over JavaScript for all new projects",
  "display_summary": "Programming language preference"
}
```

### Response
#### Success Response (200)
- **id** (string) - The unique identifier of the stored memory.

#### Response Example
```json
{
  "id": "uuid-generated-by-system"
}
```
```

--------------------------------

### Ingest Content to Memory

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Store content to memory with optional display summary and project categorization. The result indicates the ingestion status.

```typescript
const result = await client.ingest(
  "User completed the React certification course on March 1st, 2026.",
  {
    displaySummary: "React certification completed",
    project: "learning"
  }
);
console.log(`Ingestion status: ${result.status}`);
```

--------------------------------

### Update Wiki Document

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Updates an existing wiki document with new title, content, or collection. Requires the document ID.

```typescript
const updated = await client.updateWikiDocument(
  "doc-abc123",
  {
    title: "Updated Deployment Guide",
    content: "# New Content\n\nRevised steps...",
    collection: "archived"
  }
);
console.log(`Updated: ${updated.title}`);
```

--------------------------------

### Delete Memories with Confirmation

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

A two-step process to delete memories. The first step searches for memories to delete, and the second step, requiring confirmation and a UUID, performs the deletion.

```typescript
// Step 1: Search for memories to delete
const searchResult = await membase_forget({
  query: "outdated project deadline"
});
// Returns matching memories with UUIDs for user selection
```

```typescript
// Step 2: Delete after user confirms
const deleteResult = await membase_forget({
  query: "outdated project deadline",
  uuid: "550e8400-e29b-41d4-a716-446655440000",
  confirm: true
});
```

--------------------------------

### membase_update_wiki

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Updates an existing wiki document's title, content, or collection. It's recommended to use membase_search_wiki first to obtain the document ID.

```APIDOC
## membase_update_wiki

### Description
Update an existing wiki document's title, content, or collection. Use membase_search_wiki first to find the document ID.

### Request Example
```typescript
// Update document title
const updated = await membase_update_wiki({
  doc_id: "doc-abc123",
  title: "Updated API Guide v2"
});

// Update document content
const contentUpdate = await membase_update_wiki({
  doc_id: "doc-abc123",
  content: "# Revised Content\n\nNew documentation..."
});

// Move to different collection
const moveDoc = await membase_update_wiki({
  doc_id: "doc-abc123",
  collection: "archived-docs"
});

// Update multiple fields
const fullUpdate = await membase_update_wiki({
  doc_id: "doc-abc123",
  title: "Complete API Reference",
  content: "# Full Documentation\n\n...",
  collection: "api-docs"
});
```
```

--------------------------------

### membase_search

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Search stored memories by semantic similarity with support for date filtering, source filtering, and pagination. Returns episode-centric bundles containing memories with related nodes and edges from the knowledge graph.

```APIDOC
## membase_search

### Description
Search stored memories by semantic similarity with support for date filtering, source filtering, and pagination. Returns episode-centric bundles containing memories with related nodes and edges from the knowledge graph.

### Method
POST

### Endpoint
/membase/search

### Parameters
#### Request Body
- **query** (string) - Required - The search query string.
- **date_from** (string) - Optional - The start date for filtering (ISO 8601 format).
- **date_to** (string) - Optional - The end date for filtering (ISO 8601 format).
- **timezone** (string) - Optional - The timezone for date filtering (e.g., "Asia/Seoul").
- **sources** (array of strings) - Optional - Filter memories by source (e.g., ["slack", "gmail"]).
- **project** (string) - Optional - Filter memories by project name.
- **limit** (integer) - Optional - The maximum number of results to return.
- **offset** (integer) - Optional - The number of results to skip for pagination.

### Request Example
```json
{
  "query": "user preferences for dark mode",
  "limit": 20
}
```

### Response
#### Success Response (200)
- **results** (array) - An array of memory bundles.
- **results[].memories** (array) - Memories within the bundle.
- **results[].nodes** (array) - Related knowledge graph nodes.
- **results[].edges** (array) - Related knowledge graph edges.

#### Response Example
```json
{
  "results": [
    {
      "memories": [
        {
          "id": "uuid-1",
          "content": "User prefers dark mode.",
          "timestamp": "2023-10-27T10:00:00Z"
        }
      ],
      "nodes": [],
      "edges": []
    }
  ]
}
```
```

--------------------------------

### Register Auto-Recall Hook

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Registers the auto-recall hook to automatically inject relevant memories before each AI turn. Configuration options control recall behavior and context budget.

```typescript
import { registerRecallHook } from "./hooks/recall";

// Configuration options
const config = {
  autoRecall: true,      // Enable memory recall
  autoWikiRecall: true,  // Enable wiki recall
  maxRecallChars: 4000   // Context budget (500-16000)
};

// Register the hook (done automatically by plugin)
registerRecallHook(api, client, config);
```

--------------------------------

### Store Information in Long-Term Memory

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Use to save explicit user requests or proactive context like preferences, goals, and projects. Supports categorization by project.

```typescript
const result = await membase_store({
  content: "User prefers TypeScript over JavaScript for all new projects",
  display_summary: "Programming language preference"
});
```

```typescript
const projectMemory = await membase_store({
  content: "The backend API uses PostgreSQL with Prisma ORM. Connection pooling is configured for 10 max connections.",
  display_summary: "Backend database configuration",
  project: "backend-api"
});
```

```typescript
const backgroundResult = await membase_store({
  content: "User is a senior software engineer at TechCorp, specializing in distributed systems and cloud architecture. Started role in March 2024.",
  display_summary: "Professional background"
});
```

--------------------------------

### Update Wiki Document

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Updates the properties of an existing wiki document. You can modify its title, content, and collection.

```APIDOC
## POST /wiki/documents/{id}

### Description
Update an existing wiki document's properties.

### Method
POST

### Endpoint
/wiki/documents/{id}

### Parameters
#### Path Parameters
- **id** (string) - Required - The unique identifier of the wiki document to update.

#### Request Body
- **title** (string) - Optional - The new title for the wiki document.
- **content** (string) - Optional - The new content for the wiki document.
- **collection** (string) - Optional - The new collection to which the wiki document belongs.

### Request Example
```json
{
  "title": "Updated Deployment Guide",
  "content": "# New Content\n\nRevised steps...",
  "collection": "archived"
}
```

### Response
#### Success Response (200)
- **title** (string) - The updated title of the wiki document.

#### Response Example
```json
{
  "title": "Updated Deployment Guide"
}
```
```

--------------------------------

### Search Wiki Documents

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Search wiki documents with optional collection filtering. Displays the title, ID, collection name, and similarity score for each document found.

```typescript
const wikiResult = await client.searchWiki(
  "deployment guide",  // query
  10,                  // limit
  "devops"            // collection
);

for (const doc of wikiResult.documents) {
  console.log(`${doc.title} (${doc.id})`);
  console.log(`  Collection: ${doc.collection_name}`);
  console.log(`  Similarity: ${doc.similarity}`);
}
```

--------------------------------

### Delete Wiki Document

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Delete wiki documents using a two-step confirmation flow. Search for documents first, then confirm deletion.

```typescript
// Step 1: Search for documents to delete
const searchResult = await membase_delete_wiki({
  query: "outdated documentation"
});
// Returns matching documents with IDs
```

```typescript
// Step 2: Delete after confirmation
const deleteResult = await membase_delete_wiki({
  query: "outdated documentation",
  doc_id: "doc-xyz789",
  confirm: true
});
```

```typescript
// Search within collection before deleting
const collectionSearch = await membase_delete_wiki({
  query: "old API docs",
  collection_id: "archived"
});
```

--------------------------------

### membase_delete_wiki

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Deletes wiki documents through a two-step confirmation process, similar to memory deletion.

```APIDOC
## membase_delete_wiki

### Description
Delete wiki documents using a two-step confirmation flow similar to memory deletion.

### Request Example
```typescript
// Step 1: Search for documents to delete
const searchResult = await membase_delete_wiki({
  query: "outdated documentation"
});
// Returns matching documents with IDs

// Step 2: Delete after confirmation
const deleteResult = await membase_delete_wiki({
  query: "outdated documentation",
  doc_id: "doc-xyz789",
  confirm: true
});

// Search within collection before deleting
const collectionSearch = await membase_delete_wiki({
  query: "old API docs",
  collection_id: "archived"
});
```
```

--------------------------------

### Logout from Membase CLI

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Disconnect from Membase and clear stored OAuth tokens by executing the logout command.

```bash
# Clear credentials and disconnect
openclaw membase logout
```

--------------------------------

### EpisodeBundle Type Definition

Source: https://context7.com/aristoapp/openclaw-membase/llms.txt

Defines the structure of the EpisodeBundle, which is the primary response type for memory searches. It includes episode details, relevance scores, and related knowledge graph nodes and edges.

```typescript
interface EpisodeBundle {
  episode: {
    uuid: string;
    name: string;
    labels?: string[];
    summary: string | null;
    source: string | null;
    created_at: string | null;
    valid_at?: string | null;
    attributes?: Record<string, unknown>;
  };
  relevance_score?: number | null;
  nodes?: NodeResponse[];
  edges: {
    uuid: string;
    name: string;
    fact: string | null;
    source_node_uuid?: string;
    target_node_uuid?: string;
    created_at: string | null;
  }[];
}
```