PG

parksanggwon/google-flow-mcp

开发工具

本页由 AI 翻译,技术细节请以原文为准。 查看原文

9 stars 6 forks 质量 84 趋势 25MIT

Unofficial MCP server for Google Flow (Veo / Nano Banana): video & image generation plus Scene Builder automation through your own Chrome session

概览

Let your AI assistant make videos in (Veo, Nano Banana) — and stitch them into long, seamless takes with — in plain language: - "Generate an 8-second vertical clip from this keyframe: a rooster crowing on a stone wall at dawn." - "Add it to a scene and extend it twice — keep the slow push-in going, then let the camera rise into the sky." - "Make three 9:16 poster images of a copper weathervane with Nano Banana." - "Show me what the Flow page looks like right now." (flow_inspect / flow_screenshot) This is a Model Context Protocol (MCP) server that drives Google Flow through . It works with Claude Code, Claude Desktop, Cursor, VS Code, Gemini CLI — any MCP client — and with plain JSON-RPC scripts. Chrome is launched Playwright (so Google's sign-in accepts it) with a dedicated profile that keeps your login. The server never sees a password; prompts go to Flow exactly as you wrote them; downloads come back through the browser's own authenticated session.

README

以下为原文 README 节选,完整内容请查看原文。

🎬 google-flow-mcp

English | 한국어

Let your AI assistant make videos in Google Flow (Veo, Nano Banana) — and stitch them into long, seamless takes with Scene Builder — in plain language:

  • “Generate an 8-second vertical clip from this keyframe: a rooster crowing on a stone wall at dawn.”
  • “Add it to a scene and extend it twice — keep the slow push-in going, then let the camera rise into the sky.”
  • “Make three 9:16 poster images of a copper weathervane with Nano Banana.”
  • “Show me what the Flow page looks like right now.” (flow_inspect / flow_screenshot)

This is a Model Context Protocol (MCP) server that drives Google Flow through your own Chrome session. It works with Claude Code, Claude Desktop, Cursor, VS Code, Gemini CLI — any MCP client — and with plain JSON-RPC scripts.

❌ Without / ✅ With

❌ Without ✅ With google-flow-mcp
Flow has no public API; every clip is a manual click session Your agent calls flow_generate_video, waits for the render and gets the file path back
Veo clips are capped at 8 seconds flow_scene_extend adds 7-second hops with continuous motion and audio (measured seam NCC 0.99)
One wrong click spends credits Every generating tool has a dry run (auto_confirm: false) that stops right before the click
A crash mid-render loses track of the job Job records on disk let resume: true pick a running generation up instead of paying twice
UI changes break scrapers silently flow_inspect shows the agent the live UI so a changed label is a one-line fix

🏗️ Architecture

flowchart LR
    C["🤖 MCP clientClaude Code · Claude Desktop · Cursor · VS Code · Gemini CLIor any JSON-RPC script"]

    subgraph S["google-flow-mcp — runs on your machine"]
        direction TB
        T["11 toolssession · generate · media · scene builder"]
        F["Flow automationagent composer · scene timeline · media ids"]
在 GitHub 查看完整 README

安装

npx -y @park-sang-gwon/google-flow-mcp
新手?先看 mcpskill.org 教程 →

配置

{ "mcpServers": { "google-flow": { "command": "npx", "args": ["-y", "@park-sang-gwon/google-flow-mcp"] } } }