invideo Remote Agents

AI agent · Computer-use / browser agent by InVideo

Give your AI agent a professional video editor timeline via MCP.

invideo Remote Agents connect external AI assistants like Claude and ChatGPT to a professional video editor via MCP. They allow creators to automate complex editing tasks like rough cuts and color correction on a real, collaborative timeline. Choose this for a production-grade workflow that delivers editable projects rather than just rendered clips.

Agent specs

The full spec card (context, pricing per token, modalities and more) on AntSpace. Sign up free to see the full spec card.

Languages
English

Category: AI Agents

Channels
in-app, website chat, API, CLI
Platforms
Web, Windows, macOS, Linux
Deployment
Cloud
Setup
Takes about a minute to connect to an agent.
Languages
Supports 50+ languages for captions and voiceovers.
Protocol
Model Context Protocol (MCP)
  • Connects Claude, ChatGPT, and Cursor to a professional NLE timeline via MCP
  • Real-time visual feedback with agent cursors visible in the editor
  • Performs rough cuts, color correction, and audio mixing through plain language prompts
  • Exports to MP4 or editable projects for Premiere Pro, Final Cut Pro, and DaVinci Resolve

Overview

invideo Remote Agents are AI-powered computer-use agents that connect external assistants like Claude, ChatGPT, and Cursor to a professional non-linear editing (NLE) timeline. Designed for creative professionals and editors, they solve the problem of "babysitting" AI by performing explicit editing tasks such as rough cuts, color correction, and audio mixing directly within the invideo editor. The system operates via the Model Context Protocol (MCP), allowing agents to manipulate a real multitrack timeline that users can watch in real-time and take over at any frame. Unlike standard text-to-video tools, Remote Agents deliver fully editable projects that can be exported to MP4 or professional software like Premiere Pro, Final Cut Pro, and DaVinci Resolve. Setup typically takes about a minute, enabling users to direct complex video productions through plain language prompts.

Agent Connectivity

  • MCP Integration: External agents like Claude, ChatGPT, and Cursor connect to invideo's professional timeline editor via the Model Context Protocol (MCP).
  • Live Visual Feedback: The agent opens a project in invideo Editor and sends a link where users can watch every cut, title, and transition land in real-time.
  • Collaborative Editing: Users can fix a cut or keep working alongside the agent on the same timeline, taking over at any frame.

Editing Tasks

  • Rough Cuts: Agents can cut a 30-second teaser from provided footage, selecting the strongest moments and adding title cards.
  • Color Correction: Agents can color correct multiple clips to match skin tones and apply film grain across the sequence.
  • Automated Captioning: Agents add word-by-word captions with custom styling, such as bold white text on dark backings.
  • Audio Mixing: Agents can duck music under voiceover, clean up room hum, and bring the mix to -14 LUFS.
  • Media Generation: Agents can generate b-roll shots, such as a city at sunrise, and place them between existing interview segments.
  • Format Delivery: Agents can rebuild a horizontal cut as a 9:16 vertical video for Reels and export the result.

What it does

  • Rough cutting teasers from footage\nColor correcting clips for matching skin tones\nAdding word-by-word captions with custom styling\nMixing audio with ducking and hum removal\nGenerating and placing b-roll shots\nRebuilding horizontal cuts into vertical formats

Use cases

  • Editors use Claude to perform rough cuts on raw footage, identifying the strongest moments and adding title cards automatically.
  • Content creators use ChatGPT to rebuild horizontal videos into 9:16 vertical formats for Instagram Reels and TikTok.
  • Video producers use Cursor to color correct interview clips for consistent skin tones and apply film grain across a sequence.
  • Marketers use agents to add word-by-word captions and duck music under voiceovers to ensure clear audio delivery.
  • Filmmakers generate b-roll shots through the agent and place them between scenes to fill visual gaps in a project.

Specifications

Channels
in-app, website chat, API, CLI
Integrations
Claude, ChatGPT, Cursor, Claude Code, Codex, Premiere Pro, Final Cut Pro, DaVinci Resolve
Platforms
Web, Windows, macOS, Linux
Human control
Users can watch edits in real-time and take over the timeline at any frame to fix cuts or continue working.
Deployment
Cloud
Setup
Takes about a minute to connect to an agent.
Languages
Supports 50+ languages for captions and voiceovers.
Protocol
Model Context Protocol (MCP)
Supported Clients
Claude Code, Codex, Cursor, Hermes Agent, OpenClaw, Claude (web/desktop), ChatGPT (apps directory)
Server URL
mcp.invideo.io/mcp
Export Formats
MP4, MOV, WebM, AVI, Premiere Pro, Final Cut Pro, DaVinci Resolve
Setup Time
~1 minute
Language Support
50+ languages
Audio Target Level
-14 LUFS
Interface Type
Professional NLE Timeline

How it works

  1. Connect the Agent: Users add invideo to their preferred agent like Claude or ChatGPT with one click or a single prompt.
  2. Open the Editor: The agent opens a project in the invideo Editor and provides a link for the user to join.
  3. Direct the Edit: Users give instructions in plain language, such as "rough cut this teaser" or "add captions."
  4. Collaborate and Refine: The user watches the agent's cursor perform edits in real-time and can take over the timeline at any moment.
  5. Export and Deliver: Once the edit is complete, the user exports the final video or a project file for professional NLE software.

FAQ

What are Remote Agents?

Remote Agents are AI assistants from outside invideo, such as Claude or Cursor, that connect to the invideo editor over MCP to edit the same timeline as the user.

How is this different from a text-to-video tool?

Unlike text-to-video tools, Remote Agents apply explicit changes to an editable timeline, allowing users to change any edit or restore earlier checkpoints.

Can I edit while my agent is working?

Yes, users and their agents work in the same editor simultaneously, and users can see the agent's cursor and every edit as it happens.

What can I export?

Users can export a rendered video or an editable project for professional software like Premiere Pro, Final Cut Pro, or DaVinci Resolve.

Added by AntSpace from public sources (invideo.io).

Related

More AI agents by InVideo