Integration · OpenClaw runtime

OpenClaw ⇄ SwiftyClip

OpenClaw is an autonomous-agent runtime that reads MCP tool catalogs and schedules its own work. Pointing it at SwiftyClip gives you a lights-out clipper that runs overnight without any human in the loop.

1. Install SwiftyClip

Get the notarized DMG from /download. Launch once.

2. Register as an OpenClaw MCP server

In your OpenClaw config (~/.openclaw/mcp.json):

{
  "servers": {
    "swiftyclip": {
      "transport": "stdio",
      "command": "swiftyclip",
      "args": ["mcp"]
    }
  }
}

OpenClaw discovers the ten clip.* tools automatically via tools/list. It can also pull the static catalog from /api/mcp/schema ahead of time for planning.

3. Give OpenClaw a mission

Every night at midnight, check the ~/Recordings/Inbox/ folder. For each new video, ingest it into SwiftyClip, score segments, render the top 3 with the Bold caption preset, and schedule them for 9am ET the next day on TikTok. Wake me if any render fails.

OpenClaw composes that into a recurring plan that uses clip.ingest → clip.analyze → clip.scoreSegments → clip.render → clip.schedule. A single prompt configures a permanent overnight workflow.

Studio-tier pairing

Combine OpenClaw's scheduling with SwiftyClip's outbound webhooks: when a render completes, fire a POST to your Slack/Discord/Zapier pipeline so the whole team sees the new clip candidate without opening a Mac.