ClipReel for Agents

Video editing MCP server

Connect an MCP client to ClipReel to create projects, propose timeline edits, review previews, and export editable cuts.

Connect an agent

Try the edit on real footage

Compare the eight-second source shot and the six-second edit. Open the sample project with the request ready to send.

Trim the first two seconds, add the title “A day on the ridge” for three seconds, make the grade cooler, and fade out in the last half second.
Try in ClipReel Download the source clip
Original shot
Trim, title, and fade

Connect to ClipReel

Create a scoped token, save it as CLIPREEL_API_TOKEN, and configure your client. Use a project read, project write, and job token for the walkthrough.

[mcp_servers.clipreel]
url = "https://www.clipreel.com/mcp"
bearer_token_env_var = "CLIPREEL_API_TOKEN"

Start with a project and its source files

Use create_project to name a private project. import_media supplies a signed upload URL so your client can PUT the original file. inspect_media and get_timeline expose asset and revision IDs for the next edit.

Use the editing tool for the right job

Use edit_by_prompt for a multi-step request. For a precisely timed change, send ranges to trim_clips, use add_captions for timestamped cues, or pick an aspect ratio with reframe_video. Each produces a reviewable proposal tied to the project revision.

Preview, apply, and export

render_preview queues a review copy. get_job returns the state and signed video link when available. The project owner approves destructive changes in ClipReel. apply_edit saves the approved timeline. export_video accepts a saved revision and aspect ratio, which pins what the worker renders.

Use ClipReel from an OpenAI agent workflow

In a Responses API request, add ClipReel to tools as a remote MCP server and pass your scoped token in authorization. Keep it in your application’s environment. Approve the proposed changes in ClipReel before applying a footage removal.

const clipreel = {
  type: "mcp",
  server_label: "clipreel",
  server_url: "https://www.clipreel.com/mcp",
  authorization: process.env.CLIPREEL_API_TOKEN,
  require_approval: "always"
};

Client setup and workflows

Video editing in Claude Code

Ask Claude Code to turn recordings into an editable cut. Review a preview, then open the same project in ClipReel.

Claude Code setup
Trim the first two seconds.

Video editing MCP server

Connect an MCP client to ClipReel to create projects, propose timeline edits, review previews, and export editable cuts.

MCP setup

Edit product videos from Cursor

Use Cursor and ClipReel to turn product walkthrough recordings into an editable video with captions and multiple export formats.

Cursor workflow

Automate video editing with AI agents

Use the ClipReel REST API to automate product demos, social cutdowns, captions, and reviewable edits with durable render jobs.

REST API guide

Hosted jobs use your ClipReel plan. See chat edits and video credits by plan.