invideo MCP Server
Product · API / developer platform by InVideo
Give your AI agent a professional video editor via MCP.
invideo MCP Server brings professional NLE video editing to AI agents like Claude and ChatGPT. It allows developers to automate complex editing tasks on a real timeline while maintaining the ability to export to Premiere Pro or DaVinci Resolve.
Category: Developer Tools
- Platforms
- Web browsers, Windows, MacOS, Chrome OS, Linux
- Deployment
- Cloud-based remote server via Streamable HTTP
- Languages
- Supports 50+ languages for voiceovers and captions
- Protocol Support
- Model Context Protocol (MCP)
- Server URL
- mcp.invideo.io/mcp
- Communication Protocol
- Streamable HTTP
- Edit professional NLE timelines via MCP
- Real-time visual feedback with live cursors
- Export to Premiere Pro, Final Cut, and DaVinci Resolve
- Integrated with Claude, ChatGPT, and Cursor
Overview
invideo MCP Server is a developer-centric tool that exposes a professional non-linear editing (NLE) timeline to external AI agents via the Model Context Protocol (MCP). It allows businesses and creators to connect third-party assistants like Claude, ChatGPT, and Cursor directly to the invideo editor to perform complex tasks such as color correction, audio mixing, and frame-accurate trimming. Unlike standard text-to-video generators, it produces an editable multitrack timeline where users can watch their agent's edits land in real-time and take manual control at any frame. The platform supports exporting finished work to industry-standard software including Premiere Pro, Final Cut Pro, and DaVinci Resolve. By bridging the gap between LLMs and professional video production, it enables automated yet reversible video editing workflows.
Developer Integration
- MCP Standard Support: The server implements the Model Context Protocol (MCP), an open standard that allows AI agents to work with professional tools.
- CLI Integration: Developers can connect Claude Code, Codex, Cursor, and other CLI-based agents to the invideo editor with a single prompt.
- Streamable HTTP Server: The server uses Streamable HTTP at mcp.invideo.io/mcp for communication between the agent and the editor.
- Automated Project Setup: Agents can open a professional non-linear editing (NLE) timeline and send a project link directly to the user.
Editing Capabilities Galore
- Timeline Manipulation: Agents can perform frame-accurate cuts, trims, and ripple edits on a multitrack timeline.
- Color Grading Tools: The server enables agents to color correct footage, including matching skin tones across multiple clips.
- Audio Engineering: Agents can mix audio, duck music under voiceovers, and normalize levels to targets like -14 LUFS.
- Automated Captioning: The server supports adding word-by-word captions with customizable styles and positioning.
Workflow and Handoff
- Live Visual Feedback: Users can watch their agent's edits land in real-time under a cursor labeled with the agent's name.
- Collaborative Editing: Humans can take over the timeline at any frame to fix cuts or refine the agent's work manually.
- Professional Handoffs: Finished projects can be exported as rendered videos or as editable project files for Premiere Pro, Final Cut Pro, and DaVinci Resolve.
- Generative Media Access: Agents can generate new b-roll shots, images, and audio directly into the sequence using integrated AI models.
Use cases
- Developers connect Claude Code to automate rough cuts and b-roll generation for social media content.
- Editors use ChatGPT to color correct multiple interview clips simultaneously to ensure consistent skin tones.
- Marketing teams use Cursor to automatically add word-by-word captions and duck music under voiceovers.
- Production studios export agent-built timelines to Premiere Pro for final high-end finishing and delivery.
Specifications
- APIs and SDKs
- Model Context Protocol (MCP) server at mcp.invideo.io/mcp, invideo CLI
- Platforms
- Web browsers, Windows, MacOS, Chrome OS, Linux
- Deployment
- Cloud-based remote server via Streamable HTTP
- Integrations
- Claude Code, ChatGPT, Cursor, Codex, Premiere Pro, Final Cut Pro, DaVinci Resolve
- Languages
- Supports 50+ languages for voiceovers and captions
- Protocol Support
- Model Context Protocol (MCP)
- Server URL
- mcp.invideo.io/mcp
- Communication Protocol
- Streamable HTTP
- Supported CLI
- invideo CLI via GitHub
- Export Formats (Video)
- MP4, MOV, WebM, AVI
- Export Formats (Project)
- Premiere Pro, Final Cut Pro, DaVinci Resolve
- Audio Normalization Target
- -14 LUFS
- Language Support
- 50+ languages for TTS and captions
How it works
- Connect the Agent: Users select their preferred AI agent (Claude, ChatGPT, or Cursor) and use a one-click connection or a CLI prompt to initialize the link.
- Initialize MCP Server: The agent follows the setup guide at invideo.io/agents.md to register itself as an MCP server and requests the user to sign in to their invideo account.
- Open Editor and Edit: The agent opens a project in the invideo Editor and provides a link where the user can watch edits happen live on the timeline.
- Review and Export: Once the agent completes the rough cut, color grading, or audio mix, the user can export the result as a video or a professional project file.
Pricing
- Manual timeline editing
- Free
- Voice and media generation
- Uses invideo credits
FAQ
Which agents and apps work with it?
Any client supporting remote MCP servers works, including Claude Code, Codex, Cursor, Hermes Agent, OpenClaw, and the ChatGPT desktop or web apps.
Do I need an invideo account to use it?
Yes, you sign in with an invideo account during the connection process; no separate API key is required as the client prompts for sign-in.
How is this different from a text-to-video tool?
Unlike simple text-to-video tools, the agent applies explicit changes like cuts and transitions to an NLE timeline, resulting in an editable project rather than just a rendered clip.
What can I export from the MCP server?
You can export a rendered video file or an editable project for professional software like Premiere Pro, Final Cut Pro, or DaVinci Resolve.
Added by AntSpace from public sources (invideo.io).