00:00:00Product · Your agent fixes it

Your agent fixes it. You see every step.

Claude Code, Codex, Cursor or any MCP client hears a note the moment it’s sent, asks on the frame when it isn’t sure, uploads the new version and answers. Lampo shows each step as it happens.

Start free See how it works

00:00:00The note as data

Every note arrives as data, not prose.

The frame’s number at the file’s real rate, the box in the video’s own pixels, the words, the team’s playbook and the marked frame itself. Your agent has nothing to interpret and nothing to guess.

Required work first: must, should and nice, then ideas, then the questions it asked and still waits on.
The playbook and the taste before it renders: what your team decided, and what your notes taught so far.
A client’s note says so: what the agent writes on it reaches the client, so it leaves paths and tools out.
One line per event while it waits, from lampo watch or MCP: nothing costs a token until something happens.

00:00:00Questions

When it isn’t sure, it asks.

A question pinned to the frame, with the answers it expects as buttons. Or takes to audition before it spends a render: two narrators, two music beds, each at the same loudness. Your picks come back to it as one line.

00:00:00Live

You see what it’s doing, at no token cost.

Every Send is one piece of work. A line above the notes says where it stands, and the Agent tab beside them has the rest: your notes as its plan, what it is doing in its own words, the steps it took. Lampo builds it from the calls your agent makes anyway.

The line above the notes: “fixing 3 of 6”, “rendering V4 · 42 % · about 1 min left”, “needs you · a question” with Answer, “V4 is ready” with Check fixes.
The Agent tab: your notes as its plan (next, on it, fixed, asked you), what it said to you, its steps, Stop and Tell it…
A render’s progress through lampo render: the stage, the percent and the time left, from Remotion, ffmpeg, After Effects or Blender.
When it fails, waits for a permission or goes quiet, the Inbox and your phone say so, with Try again, Nudge or Stop.
No token spent on it: it comes from the calls it makes anyway, lampo or MCP, never a file’s contents.
Starting it: on Lampo Cloud a webhook can start your agent where it lives; on your own machine Lampo can resume its Claude Code session.

00:00:00V2

Every note gets an answer.

Fixed, won’t fix with a reason, or a note of its own to say what else it changed. The open notes carry over to V2, Lampo outlines what really changed since V1, and each fix waits for you: Looks right, or Still wrong.

Who made each version: “launch-edit · 2 min · 1 fixed · 1 asked” in the version picker, and its steps one click away.
Its own notes are questions or info: the severities stay your reviewers’.
Approving stays with people, and Final means done: on a final video its fixes are refused until someone reopens it.

00:00:00Any agent

Connect it once. Then one sentence.

Pick yours: the one snippet or connector address Lampo Cloud gives it, with the step that signs it in. Then tell it Use Lampo for "Spring launch": it finds the project, puts up V1 itself, works your notes, puts up the next version, marks them fixed and waits for your notes until you approve.

Connect your agent to Lampo Cloud

Its address is https://app.lampo.video/mcp. The first time your agent connects, it opens Lampo’s sign-in in your browser: sign in and allow it, with no token to copy.

In a terminal:

claude mcp add --transport http lampo https://app.lampo.video/mcp

Then sign in: run /mcp in Claude Code and follow the steps in your browser (or claude mcp login lampo).

In ~/.codex/config.toml:

[mcp_servers.lampo]
url = "https://app.lampo.video/mcp"
tool_timeout_sec = 330

Then sign in once: codex mcp login lampo.

In ~/.cursor/mcp.json (or .cursor/mcp.json in a project):

{
  "mcpServers": {
    "lampo": {
      "url": "https://app.lampo.video/mcp"
    }
  }
}

Cursor offers to connect: sign in when it asks.

In .vscode/mcp.json (or "MCP: Open User Configuration"):

{
  "servers": {
    "lampo": {
      "type": "http",
      "url": "https://app.lampo.video/mcp"
    }
  }
}

VS Code asks you to sign in the first time.

In ~/.gemini/config/mcp_config.json (or .agents/mcp_config.json in a project):

{
  "mcpServers": {
    "lampo": {
      "serverUrl": "https://app.lampo.video/mcp"
    }
  }
}

Antigravity signs in by itself the first time. Its config opens from the MCP Servers panel: ⋯ → Manage MCP Servers → View raw config.

In ~/.config/devin/mcp_config.json (Cascade panel ⋯ → Open MCP config file):

{
  "mcpServers": {
    "lampo": {
      "url": "https://app.lampo.video/mcp"
    }
  }
}

Windsurf is Devin Desktop now; it signs in the first time it uses the server.

In ~/.gemini/settings.json:

{
  "mcpServers": {
    "lampo": {
      "httpUrl": "https://app.lampo.video/mcp"
    }
  }
}

Gemini CLI signs in by itself when the server asks; /mcp auth lampo signs in again.

In ~/.config/zed/settings.json (or .zed/settings.json in a project):

{
  "context_servers": {
    "lampo": {
      "url": "https://app.lampo.video/mcp",
      "timeout": 330
    }
  }
}

Zed asks you to sign in.

In Claude on the web or the desktop: Customize → Connectors → + Add → Add custom connector. Give it a name and the address, then sign in when asked. On Team and Enterprise plans an owner adds it for the organization first.

On the web, with Plus, Pro, Business, Enterprise or Education: turn on Developer mode under Settings → Security and login, then open ChatGPT’s plugins, select + and create a developer-mode app with the address. It signs in through Lampo.

In the client's MCP config:

{
  "mcpServers": {
    "lampo": {
      "type": "http",
      "url": "https://app.lampo.video/mcp"
    }
  }
}

A client that speaks MCP over HTTP takes the address and signs in when Lampo asks.

Then tell it one sentence: Use Lampo for "Spring launch". That is the whole loop: it finds the project, puts up V1 itself, works your notes and waits for the next ones until you approve.

On your own machine or server, Settings → Connect an agent gives the same for its address, and lampo mcp config <client> prints it.

Chat apps too: Claude and ChatGPT work through the connector alone; Claude Code and Codex render with lampo render, so you watch the progress.
There from the start: a connected agent shows in the sidebar’s Agents at once, “connected”, then “ready” while it waits for your notes.
Several at once: every write is locked and atomic, so people, agents and the lampo command work side by side.
Assign agent… gives one agent a video’s notes; Copy for an agent hands the open notes with their frames to any other.
Get started at the sidebar’s foot: your first project, your agent, its V1, then your team and a review link.
Listed in awesome-mcp-servers, under multimedia processing.

Try Lampo free.

Lampo Cloud from €0, or the open‑source app on your own machine or server. Version 0.1. The source is on GitHub.

Start free Next: Check it