openclaw-manus-research

An OpenClaw plugin that adds a manus_research tool, letting your agent delegate deep research tasks to Manus AI.

Instead of browsing the web yourself, hand off complex research to Manus — competitive analysis, market research, technical deep dives, comprehensive reports — and get structured results back.

Install

openclaw plugins add dbachelder/openclaw-manus-research

Setup

1. Get a Manus API key from manus.im/app → Settings → API Integration

2. Add it to your auth profiles (~/.openclaw/agents/main/agent/auth-profiles.json):

{
  "profiles": {
    "manus:default": {
      "type": "api_key",
      "provider": "manus",
      "key": "sk-your-key-here"
    }
  }
}

3. Enable the tool for your agent in openclaw.json:

{
  "agents": {
    "list": [
      {
        "id": "main",
        "tools": {
          "allow": ["manus_research"]
        }
      }
    ]
  }
}

4. Restart OpenClaw.

Usage

Once installed, your agent has access to the manus_research tool. Just ask it to research something:

"Research the competitive landscape for Solana trading bots — compare features, fees, and user base"

The agent will:

1. Create a Manus task via the API
2. Poll until completion (up to 10 minutes by default)
3. Return the full research results including any generated files/reports

Tool Parameters

Parameter	Type	Default	Description
prompt	string	(required)	The research task. Be specific for best results.
agent_profile	enum	"manus-1.6"	"manus-1.6" (balanced), "manus-1.6-lite" (faster), "manus-1.6-max" (most capable)
max_wait_minutes	number	10	Maximum minutes to wait for task completion

Example: Batch Research

Fire off multiple research tasks by asking your agent to research several topics. Each call creates an independent Manus task that runs in parallel on their infrastructure.

How It Works

• Create: POST /v1/tasks to Manus API with your prompt and agent profile
• Poll: GET /v1/tasks/{id} every 10 seconds until status is completed
• Extract: Pulls all assistant text output and file URLs from the response
• Return: Structured result with content + metadata (task URL, credits used, shareable link)

API Reference

• Manus API Docs
• Create Task
• OpenAI SDK Compatibility

Credits

Tasks consume Manus credits based on complexity and agent profile. manus-1.6-max uses more credits but produces better results for complex research. Check your usage at manus.im/app.

License

MIT