Batchedits

Turn your OpenClaw into an autonomous video editor using BatchEdits. Use when you need to add captions, remove silences, or apply custom styles to videos. Covers creating styles, uploading local videos, processing, and checking video status directly from WhatsApp, Telegram, or the CLI.

Setup

1. Create an account at your BatchEdits instance.
2. Obtain your OAuth Client Token from the dashboard settings.
3. Provide your OAuth token to OpenClaw. You can do this securely by running:

openclaw config set env.BATCHEDITS_API_KEY client_xxxxx

(Alternatively, you can export BATCHEDITS_API_KEY=... in your shell, or add it to the .env file in the folder where you run OpenClaw)

1. Connect the BatchEdits MCP Server to your OpenClaw setup by running this in your terminal:

openclaw config set mcp.servers.batchedits '{"type": "sse", "url": "https://batchedits.com/api/mcp"}'

Auth

The MCP server uses your .env key automatically. Under the hood, any direct HTTP uploads authenticate using standard OAuth 2.0:

Authorization: Bearer <BATCHEDITS_API_KEY>

1. Get Available Actions

Use the list_actions tool.

Returns an array of action templates (e.g., remove_silence, add_captions) with their id, name, and description. You need the id to build a style.

2. Create a Style

Use the create_style tool to build a reusable video editing preset.

Arguments:

• name: e.g. "Silence Remover"
• actions: JSON string of templates (e.g., [{"id": "action_123"}])

Returns the new styleId.

3. Upload Video

Use the upload_videos tool.

Arguments:

• filePaths: Array of absolute paths (e.g., ["/path/to/local/video.mp4"])

Because video files are large binaries, this tool returns a ready-to-run curl command. You must execute the returned curl command in your terminal to upload the local file to the server. Parse the JSON output of the curl command to get the new videoId.

4. Process Video

Use the process_video tool to start the edit.

Arguments:

• videoId: The ID from step 3
• styleId: The ID from step 2

5. Check Results

Use the list_videos tool.

Returns a list of your videos with their current status: pending, processing, completed, or failed.

Recommended Workflow for Video Editing

1. Identify the local video file provided by the user (e.g., downloaded from Telegram).
2. Determine the requested edits (captions, silence removal, etc.).
3. Check get_styles to see if a matching style already exists. If not, use list_actions and create_style to make one.
4. Call upload_videos to get the upload command.
5. Execute the curl upload command and extract the videoId.
6. Call process_video to start the job on the remote server.
7. Inform the user that the video is processing.
8. Call list_videos to check when the video reaches completed status.

Tips

• Always check get_styles first. Reusing existing styles is faster than creating a new one every time!
• Uploading videos is handled via curl because OpenClaw needs to stream large local binaries to the remote server efficiently.
• You can queue multiple videos to the same style for bulk processing by repeating the upload & process steps!