# Connect cuttingfloor to Claude, ChatGPT, Codex, Cursor and more

> cuttingfloor gives the AI agent you already use real video editing tools. Stay in Claude, ChatGPT, Codex, Cursor, Kazi or any MCP app, add one server address and sign in. Steps for each app.

cuttingfloor isn’t another editor with AI inside. It gives the agent you already use the tools to edit video. Stay where you are, add one server address and sign in.

- MCP server: https://api.cuttingfloor.ai/mcp (Streamable HTTP, OAuth sign-in)
- Scripts: a key from the account page, sent as `Authorization: Bearer cf_…`
- Steps checked 2026-11-01 against each app's own help pages
- Web page: https://cuttingfloor.ai/connect

## How it works

1. **You ask your agent.** In the app you already use, say what you want: “Cut the pauses, add big captions, make it vertical.”
2. **It calls the tools.** Your agent imports the footage, reads the transcript and makes each cut, caption and mix with cuttingfloor’s editing tools.
3. **Every edit comes back with a preview.** Each edit returns a new version of the timeline and a picture of it, so your agent sees what it made and fixes what’s off. Earlier versions stay. When it’s right, it exports an MP4 and gives you the link.

## Kazi

_Web, phone, text and meetings_

1. In Kazi (app.apifunnel.ai), open Servers and add your own server.
2. Paste the address and choose Connect.
3. Sign in when the cuttingfloor page opens.
4. Or skip the menus: ask Kazi to connect the cuttingfloor MCP server at the address.

Server address:

```
https://api.cuttingfloor.ai/mcp
```

Or say to Kazi:

```
Connect the cuttingfloor MCP server at https://api.cuttingfloor.ai/mcp
```

- Free through Kazi: an account connected from Kazi gets cuttingfloor for good.

Meet Kazi at kazi.apifunnel.ai: https://kazi.apifunnel.ai

## Claude

_claude.ai on the web and the desktop app_

1. Open Customize → Connectors, click +, then Add custom connector.
2. Name it cuttingfloor and paste https://api.cuttingfloor.ai/mcp as the URL. Click Continue.
3. Choose Sign in now (or Sign in when needed). For the OAuth client, keep Use Claude’s published identity.
4. Click Add, and sign in on the page that opens.
5. In a chat, open + → Connectors and turn cuttingfloor on.

The address to paste:

```
https://api.cuttingfloor.ai/mcp
```

- Works on Free, Pro, Max, Ultra, Team and Enterprise. Free accounts get one custom connector.
- On Team and Enterprise, an owner adds it first under Organization settings → Connectors → Add → Custom → Web. Then each person clicks Connect under Customize → Connectors.
- The Claude phone apps use connectors you added on the web, but can’t add them.

Claude's own help page: https://support.claude.com/en/articles/11175166-getting-started-with-custom-connectors-using-remote-mcp

## Claude Cowork

_The Claude desktop app, web and Chrome_

1. Add cuttingfloor as a connector in Claude, as above. Cowork uses the same connectors as your Claude account.
2. Start a Cowork task and ask for the edit. It calls cuttingfloor’s tools as it works.

The address to paste:

```
https://api.cuttingfloor.ai/mcp
```

- Cowork is on Pro, Max, Ultra, Team and Enterprise.
- On Team and Enterprise, Cowork asks you to approve each step that changes something, unless an admin turns on “Allow ‘Always allow’ for connector tools”.

Claude Cowork's own help page: https://support.claude.com/en/articles/13455879-use-claude-cowork-on-team-and-enterprise-plans

## Claude Code

_Terminal_

1. Run the command below.
2. Inside Claude Code, type /mcp and sign in on the page that opens.

Run in a terminal:

```
claude mcp add --transport http cuttingfloor https://api.cuttingfloor.ai/mcp
```

For everyone on a project (.mcp.json):

```
{
  "mcpServers": {
    "cuttingfloor": {
      "type": "http",
      "url": "https://api.cuttingfloor.ai/mcp"
    }
  }
}
```

- Add --scope project to share it through the project’s .mcp.json, or --scope user for all your projects.

Claude Code's own help page: https://code.claude.com/docs/en/mcp

## ChatGPT

_chatgpt.com (add it on the web)_

1. Open chatgpt.com/plugins.
2. Click +, then Add custom MCP server.
3. Name it cuttingfloor. Under Connection, paste https://api.cuttingfloor.ai/mcp as the Server URL.
4. Set Authentication to OAuth.
5. Accept the warning, click Create as a plugin, check the tools and click Install. Sign in when asked.
6. In a chat, type @cuttingfloor and say what you want.

The Server URL:

```
https://api.cuttingfloor.ai/mcp
```

- Which plans can add a custom MCP server is up to OpenAI and your plan.
- On Business, only admins and owners can create plugins with MCP servers. On Enterprise and Edu, plugins are off until an admin allows them.
- Plugins you install work in ChatGPT on your phone too, but you add them on the web or desktop.

ChatGPT's own help page: https://developers.openai.com/apps-sdk/deploy/connect-chatgpt

## ChatGPT Work

_ChatGPT on the web and the desktop app_

1. Add cuttingfloor as a plugin in ChatGPT, as above.
2. Start a Work task and type @cuttingfloor with the edit you want.

In a Work task:

```
@cuttingfloor cut this into a 30-second vertical clip with captions
```

- Workspace admins choose which plugins Work can use.

ChatGPT Work's own help page: https://learn.chatgpt.com/docs/get-started-with-work

## Codex

_CLI, IDE extension and the ChatGPT desktop app_

1. Run the first command below to add cuttingfloor.
2. Run codex mcp login cuttingfloor and sign in in your browser.
3. In the IDE extension: gear menu → MCP servers → add, then Restart extension.
4. In the desktop app: Settings → MCP servers → Add server, choose Streamable HTTP, paste the address, Save and Restart, then click Authenticate.

Run in a terminal:

```
codex mcp add cuttingfloor --url https://api.cuttingfloor.ai/mcp
codex mcp login cuttingfloor
```

~/.codex/config.toml:

```
[mcp_servers.cuttingfloor]
url = "https://api.cuttingfloor.ai/mcp"
```

- The CLI, the IDE extension and the desktop app share ~/.codex/config.toml, so adding it once covers all three.

Codex's own help page: https://learn.chatgpt.com/docs/extend/mcp

## Cursor

_Desktop editor_

1. Click Add to Cursor below, or paste the config into ~/.cursor/mcp.json (or .cursor/mcp.json in a project).
2. Cursor opens the sign-in the first time it uses a tool.

[Add to Cursor](cursor://anysphere.cursor-deeplink/mcp/install?name=cuttingfloor&config=eyJ1cmwiOiJodHRwczovL2FwaS5jdXR0aW5nZmxvb3IuYWkvbWNwIn0%3D)

~/.cursor/mcp.json:

```
{
  "mcpServers": {
    "cuttingfloor": {
      "url": "https://api.cuttingfloor.ai/mcp"
    }
  }
}
```

- Team admins can share it with everyone under Dashboard → Plugins & MCPs.

Cursor's own help page: https://cursor.com/docs/context/mcp

## VS Code / GitHub Copilot

_VS Code with Copilot_

1. Open the Command Palette and run MCP: Add Server.
2. Choose HTTP, paste the address, name it cuttingfloor and pick Global.
3. Start it and sign in when VS Code asks.

Or run in a terminal:

```
code --add-mcp '{"name":"cuttingfloor","type":"http","url":"https://api.cuttingfloor.ai/mcp"}'
```

.vscode/mcp.json:

```
{
  "servers": {
    "cuttingfloor": {
      "type": "http",
      "url": "https://api.cuttingfloor.ai/mcp"
    }
  }
}
```

- On Copilot Business and Enterprise, an admin has to turn on MCP servers in Copilot first (Settings → Copilot → Policies).

VS Code / GitHub Copilot's own help page: https://code.visualstudio.com/docs/copilot/customization/mcp-servers

## Antigravity CLI (Gemini)

_Terminal and the Antigravity editor_

1. Add cuttingfloor to ~/.gemini/config/mcp_config.json (or .agents/mcp_config.json in a project).
2. Run /mcp to check it’s there, and sign in when asked.
3. In the Antigravity editor: … → MCP Servers → Manage MCP Servers → View raw config, and add the same lines.

~/.gemini/config/mcp_config.json:

```
{
  "mcpServers": {
    "cuttingfloor": {
      "serverUrl": "https://api.cuttingfloor.ai/mcp"
    }
  }
}
```

- Antigravity CLI replaced Gemini CLI in June 2026. It uses serverUrl, not httpUrl.

Antigravity CLI (Gemini)'s own help page: https://antigravity.google/docs/mcp?tab=cli

## Zed

_Desktop editor_

1. Open Settings → AI → MCP Servers → Add Server → Add Remote Server, and paste the address.
2. Or add the lines below to your settings.json.
3. Zed opens the sign-in for you.

settings.json:

```
{
  "context_servers": {
    "cuttingfloor": {
      "url": "https://api.cuttingfloor.ai/mcp"
    }
  }
}
```

Zed's own help page: https://zed.dev/docs/ai/mcp

## Goose

_Desktop app and CLI_

1. Click Add to Goose below, or in the desktop app open Extensions → Add custom extension.
2. In the CLI, run goose configure → Add Extension → Remote Extension (Streamable HTTP), and paste the address.
3. Goose opens the sign-in for you.

[Add to Goose](goose://extension?url=https%3A%2F%2Fapi.cuttingfloor.ai%2Fmcp&type=streamable_http&id=cuttingfloor&name=cuttingfloor&description=Headless%20video%20editor)

~/.config/goose/config.yaml:

```
extensions:
  cuttingfloor:
    name: cuttingfloor
    type: streamable_http
    uri: https://api.cuttingfloor.ai/mcp
    enabled: true
    timeout: 300
```

Goose's own help page: https://goose-docs.ai/docs/getting-started/using-extensions

## Any MCP client

_Anything that takes a remote MCP server_

1. Add a remote server at https://api.cuttingfloor.ai/mcp. It speaks Streamable HTTP.
2. Your app opens the cuttingfloor sign-in page (OAuth). Sign in, and it’s connected.
3. For a script with no browser, make a key on your account page (it starts with cf_) and send it as a header instead.

Most apps (mcpServers):

```
{
  "mcpServers": {
    "cuttingfloor": {
      "url": "https://api.cuttingfloor.ai/mcp"
    }
  }
}
```

A script, with a key from your account page:

```
{
  "mcpServers": {
    "cuttingfloor": {
      "url": "https://api.cuttingfloor.ai/mcp",
      "headers": {
        "Authorization": "Bearer cf_…"
      }
    }
  }
}
```

## Questions

### How do I use Claude to edit video?

Add cuttingfloor as a custom connector: in Claude, open Customize → Connectors, click +, choose Add custom connector, paste https://api.cuttingfloor.ai/mcp and sign in. Then ask for the edit in a chat or a Cowork task. In Claude Code, run claude mcp add --transport http cuttingfloor https://api.cuttingfloor.ai/mcp, then /mcp to sign in.

### Can ChatGPT edit video?

Yes, with cuttingfloor added as a plugin. Open chatgpt.com/plugins, click +, choose Add custom MCP server, paste https://api.cuttingfloor.ai/mcp and set authentication to OAuth. Then type @cuttingfloor in a chat or a ChatGPT Work task.

### Is there a video editing MCP server for Codex?

cuttingfloor is one. Run codex mcp add cuttingfloor --url https://api.cuttingfloor.ai/mcp, then codex mcp login cuttingfloor to sign in. The Codex CLI, the IDE extension and the ChatGPT desktop app share the same settings, so adding it once covers all three.

### Do I have to install anything?

No. The editing and rendering run on our servers, and your agent connects to one address. If you’d rather run it on your own computer, it’s free and open source: see Run it yourself.

### Can a script use it without a browser sign-in?

Yes. Make a key on your account page (it starts with cf_) and have the script send it as Authorization: Bearer cf_… to the same address.