Secured by your API key
Every connector request is tied to the Editly key you create in Settings.
Editly MCP Server is a connector that lets Claude, Cursor, Windsurf, and other AI apps call Editly tools instead of just giving advice about video editing.
Every connector request is tied to the Editly key you create in Settings.
Works with Claude remote connectors and MCP-compatible local clients.
Your AI app turns your request into Editly tool calls and checks the render until it is done.
For claude.ai, use the hosted remote connector. For Claude Desktop, Cursor, or other local clients, the same MCP server can run via stdio.
Open Settings, go to API Keys, create a key, and copy it while it is visible.
Use the remote MCP endpoint and put your Editly key in the Authorization header.
Ask for the edit you want. Your AI app calls Editly, tracks the render, and returns the output.
Remote connector
URL: https://editly.in/api/agent. Header: Authorization: Bearer editly_your_key_here.
Tell your AI app the source video and the desired result. Keep the prompt specific about duration, format, captions, music, and pacing.
"Remove silences, add clean white captions, and keep the final video under 5 minutes."
"Crop this to 9:16 for TikTok, keep the strongest hook first, and add warm color grading."
"Find calm piano music, add it softly under the voice, and export a polished interview clip."
The practical details for setup, hosting, file transfer, and access.
No. Use the remote connector URL in Claude settings. Local installs are only for Claude Desktop, Cursor, Windsurf, Zed, or other stdio-based MCP clients.
Use an Editly-uploaded asset, a direct video URL, or a signed storage URL that stays valid long enough for the render to finish.
Only through the tools authorized by your Editly API key. You can revoke a key any time from Settings.
Create a key, add the connector, and keep video transfer on the storage/render path where it belongs.