autoMate

🤖 API Tool Center + Desktop Automation — One MCP, 31 Platforms

中文 | 日本語

The only MCP server that combines 31 platform integrations + desktop GUI automation — local-first, zero cloud dependency, your credentials never leave your machine

autoMate is an MCP server that works in two modes:

Mode 1 — API Tool Center: Set environment variables for the platforms you use — autoMate auto-registers native tools for each one. Send messages, create issues, manage tasks, process payments, search contacts — all from one MCP.

Mode 2 — Desktop GUI Automation: Give Claude hands and eyes to control any desktop application with no API — 剪映, Photoshop, AutoCAD, SAP, internal tools.

autoMate's unique position: The only MCP server that is simultaneously local-first, Chinese-platform-native, open-source, and combines API integrations with desktop GUI automation.

Prerequisite: pip install uv

Open Settings → Developer → Edit Config, then add:

{
  "mcpServers": {
    "automate": {
      "command": "uvx",
      "args": ["automate-mcp@latest"]
    }
  }
}

Restart Claude Desktop — done. @latest keeps autoMate up to date automatically.

Add env vars for whichever platforms you use:

{
  "mcpServers": {
    "automate": {
      "command": "uvx",
      "args": ["automate-mcp@latest"],
      "env": {
        "SLACK_BOT_TOKEN": "xoxb-...",
        "GITHUB_TOKEN": "ghp_...",
        "NOTION_API_KEY": "secret_...",
        "STRIPE_SECRET_KEY": "sk_live_...",
        "FEISHU_APP_ID": "cli_...",
        "FEISHU_APP_SECRET": "...",
        "SENTRY_AUTH_TOKEN": "...",
        "SENTRY_ORG_SLUG": "my-org"
      }
    }
  }
}

Only set the ones you need — unused integrations are silently skipped, zero overhead.

Settings → MCP Servers → Add:

{
  "automate": {
    "command": "uvx",
    "args": ["automate-mcp@latest"]
  }
}

Script library — save once, run forever:

Cloud Vision — autonomous UI understanding (requires HF config):

Low-level desktop control:

Add HuggingFace env vars to enable autonomous screen parsing and action reasoning:

"env": {
  "AUTOMATE_HF_TOKEN": "hf_...",
  "AUTOMATE_SCREEN_PARSER_URL": "https://your-omniparser-endpoint.aws.endpoints.huggingface.cloud",
  "AUTOMATE_ACTION_MODEL_URL": "https://your-uitars-endpoint.aws.endpoints.huggingface.cloud",
  "AUTOMATE_ACTION_MODEL_NAME": "ByteDance-Seed/UI-TARS-1.5-7B",
  "AUTOMATE_HF_NAMESPACE": "your-hf-username",
  "AUTOMATE_SCREEN_PARSER_ENDPOINT": "omniparser-v2",
  "AUTOMATE_ACTION_MODEL_ENDPOINT": "ui-tars-1-5-7b"
}

See .env.example for the full reference.

Scripts are saved as .md files in ~/.automate/scripts/ — human-readable, git-friendly, shareable.

---
name: jianying_export_douyin
description: Export the current 剪映 project as a 9:16 Douyin video
created: 2025-01-01
---

## Steps

1. Open export dialog [key:ctrl+e]
2. Select resolution 1080×1920 [click:coord=320,480]
3. Set format to MP4 [click:coord=320,560]
4. Click export [click:coord=800,650]
5. Wait for export to finish [wait:5]

Q: Which integrations are active?
Only integrations whose env vars are all set. Unset ones are silently skipped — no errors, no overhead.

Q: How is this different from Composio or Zapier MCP?
Composio and Zapier route your API calls through their cloud servers and require account creation. autoMate runs entirely on your machine — your credentials stay local. autoMate also covers Chinese platforms (Feishu, DingTalk, WeChat, etc.) that no cloud-hosted MCP service supports, and bundles desktop GUI automation for apps with no API.

Q: Do I need a GPU for Cloud Vision?
No — everything runs on HuggingFace Inference Endpoints. You only need a HF token and deployed endpoints.

Q: Does it work on macOS / Linux?
Yes — all three platforms are supported.