Skip to main content

Integrating with OpenClaw, OpenCode & More

MoLOS provides 72+ MCP tools that give your AI coding assistants permanent memory and the ability to manage your tasks, projects, and knowledge base.

Quick Setup​

  1. Generate an MCP Auth Key in MoLOS
  2. Add MoLOS MCP server to your AI coding assistant
  3. Update system prompt to enable tool usage

That's it! Your AI can now read and update your MoLOS data.

Step 1: Generate MCP Auth Key​

In MoLOS:​

  1. Navigate to AI → MCP
  2. Click Create New Key
  3. Name it (e.g., "OpenClaw", "Cursor", "OpenCode")
  4. Choose permissions (which modules the AI can access)
  5. Click Generate
  6. Copy the key immediately (you won't see it again)

MCP Auth Keys

tip

Create separate keys for each AI tool. This way you can revoke access individually if needed.

Key Permissions​

Permission ModuleWhat AI Can Do
TasksRead/create/update tasks, projects, and areas
MarkdownAccess documents and quick notes
GoalsView and update goals and progress
HealthLog and retrieve health entries
LLM CouncilRun consultations and view history
warning

Only enable modules you want the AI to access. For coding assistants, Tasks and Markdown are usually sufficient.

Step 2: Add MoLOS to Your AI Tool​

OpenClaw​

Add MCP Server:

  1. Open OpenClaw settings
  2. Navigate to MCP Servers
  3. Click Add Server
  4. Configure with:
{
  "name": "MoLOS",
  "url": "http://localhost:4173/mcp",
  "headers": {
    "Authorization": "Bearer YOUR_AUTH_KEY_HERE"
  }
}
  1. Click Save
  2. Restart OpenClaw if needed

OpenCode​

Configure MCP:

  1. Open OpenCode settings
  2. Go to Extensions → MCP
  3. Click Add MCP Server
  4. Enter:
    • Name: MoLOS
    • URL: http://localhost:4173/mcp
    • API Key: YOUR_AUTH_KEY_HERE
  5. Click Connect

Cursor​

Setup MCP:

  1. Open Cursor settings (Cmd+,)
  2. Navigate to Model Context Protocol
  3. Click Add Server
  4. Configure:
    • Server Name: MoLOS
    • Endpoint: http://localhost:4173/mcp
    • Authorization Token: YOUR_AUTH_KEY_HERE
  5. Click Save

Windsurf​

Connect MCP:

  1. Open Windsurf settings
  2. Go to MCP Integration
  3. Add new server:
    • Name: MoLOS
    • Base URL: http://localhost:4173/mcp
    • Auth Header: Bearer YOUR_AUTH_KEY_HERE
  4. Click Connect

Other MCP-Clients​

For any MCP-compatible client, you'll need:

ParameterValue
Server URLhttp://localhost:4173/mcp
Auth TypeBearer token
Auth HeaderAuthorization: Bearer YOUR_AUTH_KEY_HERE
note

If MoLOS is running on a different port or remote server, replace localhost:4173 with your actual URL.

Step 3: Configure System Prompt​

Your AI coding assistant needs instructions on when and how to use MoLOS tools. Add this to your system prompt:

System Prompt Template​

You have access to MoLOS, a structured memory and productivity system with 72+ MCP tools.

## When to Use MoLOS Tools

Use MoLOS tools when user requests involve:

- **Task Management**: Creating, updating, or reviewing tasks and projects
- **Documentation**: Creating or searching documentation and notes
- **Project Context**: Understanding project structure, goals, or progress
- **Workflows**: Managing daily routines, habits, or productivity systems

## How to Use MoLOS Tools

1. **Identify the right tool**: Based on what user wants to do
2. **Gather parameters**: Ask for missing required information
3. **Execute the tool**: Call the appropriate MCP tool
4. **Confirm action**: Tell user what was done
5. **Offer follow-up**: Suggest related actions

## Available Tool Categories

- **Tasks Module**: Create, read, update tasks; manage projects and areas
- **Markdown Module**: Create pages; manage notes; search documents
- **Goals Module**: Track progress; set milestones
- **Health Module**: Log entries; view trends

## Example Interactions

User: "Create a task to review the PR I'm working on"
Your Response: [Uses molos_MoLOS-Tasks_bulk_create_tasks] "Created task 'Review PR' in your current project. Would you like me to add subtasks?"

User: "What documentation exists for the authentication module?"
Your Response: [Uses molos_MoLOS-Markdown_search_markdown_pages] "Found 3 pages: 'Authentication Overview', 'API Reference', 'Security Guide'"

User: "How am I doing on my goals?"
Your Response: [Uses molos_MoLOS-Goals_get_goals] "Progress: 3 of 5 goals on track (60%). Behind schedule: 1 goal"

## Best Practices

- Always confirm with user before making changes (creating tasks, updating goals)
- Ask for missing required parameters (e.g., "Which project should this task belong to?")
- Provide clear summaries of what was done
- Suggest follow-up actions when helpful

## Tool Discovery

To see all available tools and their parameters, ask MoLOS to show you the tools browser or describe what you want to do and let me find the right tool.

Testing the Integration​

Verify Connection​

Test that your AI can access MoLOS tools:

Prompt:

List all available MoLOS MCP tools.

Expected Response: Your AI should list tools like:

  • molos_MoLOS-Tasks_get_tasks
  • molos_MoLOS-Tasks_bulk_create_tasks
  • molos_MoLOS-Markdown_search_markdown_pages
  • molos_MoLOS-Markdown_create_markdown_page

Test Tool Usage​

Prompt:

Create a task to "Review project documentation" with high priority.

Expected Response: AI should:

  1. Ask which project to add task to (if you have multiple)
  2. Create the task using MCP tool
  3. Confirm the task was created
  4. Show task details (title, priority, project)

Common Workflows​

Development Workflow​

You: "I'm building a new feature for the authentication module"

AI: [Uses MoLOS to check if task exists]
"Found 3 existing tasks for 'Authentication Module'. 
Would you like me to create a new subtask or add to existing?"

You: "Add a subtask to 'Implement OAuth'"

AI: [Uses molos_MoLOS-Tasks_bulk_create_tasks]
"Done! Added subtask to 'Implement OAuth'. 
Would you like me to create documentation for this task?"

Documentation Workflow​

You: "Document the API endpoints I just created"

AI: [Uses molos_MoLOS-Markdown_search_markdown_pages]
"Found existing page 'API Reference'. 
Should I append to that or create a new page?"

You: "Append to it"

AI: [Uses molos_MoLOS-Markdown_update_markdown_page]
"Added new endpoints to 'API Reference'. 
Created 4 links from code to documentation."

Review Workflow​

You: "What should I work on today?"

AI: [Uses molos_MoLOS-Tasks_get_tasks and molos_MoLOS-Goals_get_goals]
"Today's priorities:
• Review PR #42 (High priority, due today)
• Fix login bug (Urgent)
• Document API changes (Medium)

You have 3 tasks due today. Want me to help you start with the urgent one?"

Troubleshooting​

IssueSolution
AI can't connect to MoLOS?Verify MoLOS is running (docker ps) and accessible at http://localhost:4173
"Connection refused" error?Check firewall settings; ensure port 4173 is open
"Unauthorized" error?Verify auth key is correct and not expired
AI doesn't see tools?Restart your AI assistant after adding MCP server
Tools fail to execute?Check key permissions include required modules
MoLOS on remote server?Update server URL from localhost to your server address

System Prompt Variations​

For Cursor (Code-Focused)​

You have access to MoLOS MCP tools for task and knowledge management.

When coding:
1. Check MoLOS for existing tasks related to current file/project
2. Create tasks for bugs, features, or improvements you identify
3. Offer to document code changes in MoLOS markdown pages

Always ask before creating tasks or documentation.

For OpenClaw (General Purpose)​

MoLOS provides structured memory for tasks, notes, and goals.

Use MoLOS tools to:
- Track work tasks and projects
- Save research and documentation
- Reference project context when coding

Ask user for confirmation before making changes.

For Windsurf (Agile Development)​


MoLOS is your task and project management system via MCP.

When working on code:
- Update task status when features are implemented
- Create new tasks for bugs or improvements
- Reference project goals when making decisions

Always confirm before creating or updating tasks.

Advanced: Custom Tool Workflows​

Create Task from Code Comment​

Prompt for AI:

When you see TODO comments like "// TODO: Add error handling",
ask me if I want to create a MoLOS task for it.

AI Behavior:

  • Scans code for TODOs
  • Proposes creating tasks
  • Links tasks to current project
  • Tracks TODO completion

Document Function Calls​

Prompt for AI:

After you help me implement a function, offer to create MoLOS documentation for it.

AI Behavior:

  • Generates markdown documentation
  • Creates page in appropriate module
  • Links code to documentation

PR Review Integration​

Prompt for AI:

When reviewing a PR, check MoLOS for related tasks and update their status.

AI Behavior:

  • Finds tasks linked to PR
  • Updates status to "in progress" or "done"
  • Creates subtasks for review findings

Security Best Practices​

Keep Keys Secure
  • Never share auth keys in code repositories
  • Use separate keys for each AI tool
  • Rotate keys every 90 days
  • Revoke keys immediately if compromised
Least Privilege

Only enable modules the AI actually needs:

  • Coding assistants: Tasks, Markdown
  • Productivity agents: Tasks, Goals, Health, Markdown
  • Research agents: Markdown, Goals

Last Updated: March 21, 2026 | Version 1.1.0