Google Flow API MCPサーバー

Google Flow API examples (useapi.net)

English · 中文 · 日本語 · Português · Tiếng Việt · Bahasa Indonesia · ไทย

Runnable Node.js and Python examples for the Google Flow API by useapi.net — generate Veo 3.1 Fast, Quality and Lite video, Omni 1.2 and 1.1 Flash audio-native video, and Nano Banana Pro, 2.1 and 2 Lite images through a REST API that runs on your Google AI subscription for Google Flow: Free, Plus $4.99, Pro $19.99, Ultra $99.99 or Ultra $199.99 a month (as of November 2026). No Google Cloud project, no Gemini API key, no per-call metering.

Each example ships JavaScript and Python implementations (.mjs and .py) driven by a prompts.json (or, for omni-extend/, a story.json) you edit. Most are batch runners: they submit every prompt in the file and download every result, so you can queue a batch and come back to the winners. ugc-product-video/ is a pipeline instead — it chains six endpoints into one finished video, checkpointing as it goes.

Example	What it does	Tutorial	Tutorial date
demo/	Live agent chat app: a browser chat with Google Flow's creative agent through the API, with answers streamed live, the agent's images and videos inline (full-size view, download, reuse), uploads, characters and voices attached along the way, spending approvals, a Library and the agent's settings. Open the live demo or open demo/index.html locally. Vanilla JS, no build	Agent overview, Chat with Google Flow Agent	October 16, 2026
veo-video/	Batch-generate Veo 3.1 video — text-to-video, first/last-frame image-to-video	Generate Veo 3.1 video via curl · 中文	June 15, 2026 (October 20, 2026)
images/	Batch-generate images with Nano Banana 2 Lite (default), Nano Banana 2.1, Nano Banana Pro	Generate images via curl	June 15, 2026 (October 20, 2026)
nano-banana-compare/	Run one prompt through all three Nano Banana models (2 Lite, 2.1, Pro) and compare	Nano Banana 2 Lite vs 2.1 vs Pro compared	July 2, 2026 (October 20, 2026)
omni-flash/	Batch-generate Omni 1.1 Flash (or Omni 1.2 Flash with "model": "omni-flash-1.2") audio-native video — text-to-video, first/last-frame image-to-video, reference-to-video, video-to-video edit	Generate Omni 1.1 Flash video via curl	June 15, 2026 (November 6, 2026)
omni-extend/	Make one long Omni 1.2 Flash video: a first clip of 3 to 15 seconds, then extensions of up to 10 seconds each with their own voices, images or characters (the release post's 30-second scene is the default story)	Extend a video with Omni 1.2	June 15, 2026 (November 6, 2026)
ugc-product-video/	Build a whole UGC product video — product sheets, a presenter, one still, an Omni 1.1 Flash clip per product, upscaled and joined	Make a UGC product video	September 4, 2026 (October 20, 2026)
n8n/	Ready-to-import n8n workflows: a Google Sheet batch of Veo 3.1 videos and Nano Banana Pro images saved to Google Drive, and a one-prompt form	n8n workflows	September 28, 2026 (October 28, 2026)
n8n/ugc-ad-factory.json	Ready-to-import n8n form workflow that builds a whole UGC video ad with Omni 1.1 Flash: pick or re-roll at every step, upload your own product photos, 1–5 products, 1080p	Make UGC video ads in n8n	October 3, 2026 (October 20, 2026)
skills/omni-flash-video/	Agent skill for Claude Code, Codex and Cursor: ask for a clip and the agent makes it with Omni 1.2 Flash, Omni 1.1 Flash or Veo 3.1 (dialogue, voices, reference images, first/last frames) and downloads the .mp4. Install: npx skills add useapi/google-flow-api --skill omni-flash-video	SKILL.md

Quick start

You need Node.js v21 or newer or Python 3.x (neither has any dependencies to install), a useapi.net API token, and a connected Google Flow account (one $15/month subscription covers every useapi.net API and 3 Google accounts):

git clone https://github.com/useapi/google-flow-api.git
cd google-flow-api/veo-video
node ./google-flow.mjs <API_TOKEN> <EMAIL>
# or, equivalently, with Python:
python3 ./google-flow.py <API_TOKEN> <EMAIL>

Edit prompts.json in each folder to queue your own prompts. Every supported parameter is documented on the POST /videos and POST /images endpoint pages.

MCP server

useapi.net also runs a hosted Model Context Protocol server for Google Flow, so an AI assistant such as Claude Code, Cursor or VS Code Copilot can use it on your Google Flow account. It has 12 tools: chat with Google Flow Agent, attach images, a video, voices and characters, answer its requests to spend credits, manage chats, set the project's approval mode, instructions and default models, list the account's characters, voices, images and videos, and upload images. Nothing to install or run locally.

• Endpoint: https://api.useapi.net/v1/google-flow/mcp
• Transport: Streamable HTTP
• Authorization: your useapi.net API token in the Authorization: Bearer … header, and an account connected with the Google Flow setup
• Listed in the official MCP Registry as net.useapi/google-flow

Claude Code:

claude mcp add --transport http useapi-google-flow https://api.useapi.net/v1/google-flow/mcp --header "Authorization: Bearer YOUR_API_TOKEN"

Cursor, VS Code, Windsurf and the full tool list: Google Flow MCP server docs. useapi.net is a third-party API, not affiliated with Google.

Use with Claude Code

This repo is also a Claude Code plugin marketplace. The plugin adds the hosted Google Flow MCP server and the omni-flash-video agent skill, and reads your token from the USEAPI_TOKEN environment variable, so set it before you start Claude Code:

export USEAPI_TOKEN=user:12345-...      # your useapi.net API token

Then, inside Claude Code:

/plugin marketplace add useapi/google-flow-api
/plugin install google-flow@useapi-google-flow

On Claude Code 2.1.275 or later, /plugin install google-flow --marketplace useapi/google-flow-api does both in one step. Run /mcp to check that useapi-google-flow is connected.

Use with Gemini CLI

This repo is also a Gemini CLI extension with the same MCP server and the omni-flash-video agent skill:

gemini extensions install https://github.com/useapi/google-flow-api

The install asks for your useapi.net API token and keeps it in the system keychain. Change it later with gemini extensions config useapi-google-flow, and check the connection with gemini mcp list. The omni-flash-video skill's scripts read USEAPI_TOKEN from your shell, so export it as well if you use the skill.

Common questions

• Does Google Flow have an API or an API key? Through useapi.net, yes. Google offers Flow only as a website, with no public API and no Flow API key (as of November 2026). The useapi.net Google Flow API runs Veo 3.1, Omni 1.1 Flash and Nano Banana Pro, 2.1 and 2 Lite over REST on your Google AI subscription, with a useapi.net API token as the key, and also chats with Google Flow Agent, which Google offers only inside the Flow website.
• How do I connect my Google account? Through the useapi.net automated browser setup: sign in with Google once, with no cookies to copy and no Google Cloud project or Gemini API key, and it's done in about a minute. Captchas are solved for you: your first account gets 300 free, then your own captcha provider costs about $0.80 to $3.00 per 1,000. To connect by hand, the Google Account Setup scripts (Windows, macOS, Linux) open a clean, single-use browser profile, so the cookies you copy keep working.
• How much does the Google Flow API cost? Through useapi.net, Google Flow video costs a fraction of Google's own APIs. On Google AI Ultra $199.99, Omni 1.2 Flash is about $0.024 a second at 720p ($0.19 for 8 seconds), against $0.07 a second on Google Cloud, and an 8-second Veo 3.1 Fast video is $0.08, against $0.80 on Google's Agent Platform. Nano Banana Pro images use no credits, against $0.134 each (as of November 2026). The price table shows every model on every subscription.
• Is there a free Google Flow API? Through useapi.net, yes: a free Google account makes Nano Banana images, Veo 3.1 Fast and Lite videos and Omni 1.2 Flash videos of up to 10 seconds on its 50 daily Flow credits, enough for 2 Fast or 5 Lite videos, while Google's own APIs charges for every image and every second of video (as of November 2026). Upscaling an image to 2K is free on every account too. Omni 1.1 Flash, Omni 1.2 Flash longer than 10 seconds, extending with Omni 1.2 and video upscaling need a paid Google AI subscription.
• Is there a second Omni allowance? Through useapi.net, yes: the Google Vids API makes Omni 1.1 Flash video on the same Google AI subscription from Vids' own monthly allowance, separate from Flow credits and with no captcha. On Ultra $199.99 the two allowances together come to about 2,666 ten-second Omni clips a month on one Google account. Examples: useapi/google-vids-api.
• More answers: Google Flow API questions.

More articles

• Can You Use Your Google AI Subscription Through an API? What Plus, Pro and Ultra Get You · Also in 中文 · 日本語 · Português · Tiếng Việt · Bahasa Indonesia · ไทย
• How to Use the Gemini Omni 1.2 Flash API: 15-Second Videos in Google Flow and Google Vids · Also in 中文
• Veo 3.1 API Pricing: Official Google Veo API vs Google Flow API
• Google Flow Errors Explained: Upscaling Failed, PUBLIC_ERROR_UNUSUAL_ACTIVITY and Unsafe Generation
• Sora 2 API Alternatives Compared, With Code: Seedance 2.5, Omni Flash, MiniMax H3, Kling v3
• All articles: useapi.net articles

中文说明

这是 useapi.net 托管的第三方 Google Flow API 示例代码（Node.js 与 Python）。通过 REST API 调用 Veo 3.1 Fast、Quality、Lite 视频，Omni 1.1 Flash 带音频视频，以及 Nano Banana Pro、2.1 和 2 Lite 图片生成。

• 使用你的 Google AI 订阅（免费版、Plus $4.99、Pro $19.99、Ultra $99.99 或 Ultra $199.99 每月，2026 年 11 月价格），已支持新域名 flow.google.com。
• 无需 Google Cloud 项目或 Gemini API Key，无需自己部署服务或维护浏览器。
• 账号通过自动化浏览器流程绑定，登录 Google 即可，无需手动复制 Cookie。
• reCAPTCHA 由服务端自动处理（首个账号赠送 300 次免费打码，之后接入你自己的打码服务）。
• useapi.net 每月 15 美元，含所有 API，每个订阅可连接 3 个 Google 账号。
• 成本：在 Google AI Ultra $199.99 上，一条 8 秒 Veo 3.1 Fast 视频约 $0.08 的 Flow 积分，Google 官方 API 为 $0.80（2026 年 11 月）。图片生成不消耗积分，免费 Google 账号也可使用。

中文教程（Veo 3.1、Omni Flash 1.2 和 1.1、Nano Banana Pro，含价格对比与示例）：如何通过 Google Flow API 调用 Veo 3.1、Omni Flash（1.2 和 1.1）和 Nano Banana Pro。快速开始见上方 Quick start，完整文档：Google Flow API。

日本語の説明

useapi.net がホストするサードパーティの Google Flow API の、Node.js と Python のサンプルです。REST API で、Veo 3.1 Fast、Quality、Lite の動画、音声付きの Omni 1.2 Flash と 1.1 Flash の動画、Nano Banana Pro、2.1、2 Lite の画像を生成します。

• お使いの Google AI のプランで動きます：無料、Plus $4.99、Pro $19.99、Ultra $99.99、Ultra $199.99（月額、2026年11月時点）。新しいドメイン flow.google.com にも対応しています。
• Google Cloud プロジェクトも Gemini API キーも不要で、サーバーやブラウザーを自分で管理する必要もありません。
• Google アカウントは、自動化されたブラウザーの手順で接続します。Google に一度ログインするだけで、Cookie をコピーする必要はありません。
• reCAPTCHA はサーバー側で解決します。最初のアカウントには300回の無料解決が付き、その後はご自身の CAPTCHA プロバイダーを使います。
• コスト：Google AI Ultra $199.99 では、8秒の Veo 3.1 Fast の動画1本が Flow クレジットで約 $0.08 で、Google の公式 API では $0.80 です（2026年11月時点）。画像はクレジットを使わず、無料の Google アカウントでも作れます。useapi.net は月 $15 で、すべての API と、サブスクリプションごとに3つの Google アカウントが含まれます。
• サンプルでは、prompts.json から動画と画像を一括生成し、Omni 1.2 の延長で長いシーンを組み立て、UGC の商品動画を最初から最後まで作ります。

日本語の記事：Veo 3.1 API の使い方：Google Flow API と curl で動画を生成する、Google AI のサブスクは API で使える？Plus・Pro・Ultra で毎月できること。ドキュメント全体：Google Flow API。

Sobre este repositório

Exemplos em Node.js e Python da Google Flow API, uma API de terceiros hospedada pela useapi.net, para gerar por REST vídeos Veo 3.1 Fast, Quality e Lite, vídeos Omni 1.2 e 1.1 Flash com áudio, e imagens Nano Banana Pro, 2.1 e 2 Lite.

• Roda na sua assinatura do Google AI: gratuita, Plus $4,99, Pro $19,99, Ultra $99,99 ou Ultra $199,99 por mês (preços de novembro de 2026), já no novo domínio flow.google.com.
• Sem projeto no Google Cloud e sem chave da Gemini API, sem servidor nem navegador para manter.
• A conta Google é conectada por um fluxo automatizado no navegador: você entra no Google uma vez, sem copiar cookies.
• O reCAPTCHA é resolvido pelo servidor: a primeira conta ganha 300 resoluções grátis, depois você usa o seu próprio provedor de captcha.
• Custo: no Google AI Ultra $199,99, um vídeo Veo 3.1 Fast de 8 segundos gasta cerca de $0,08 em créditos do Flow, contra $0,80 na API oficial do Google (novembro de 2026). Imagens não gastam créditos e funcionam em contas Google gratuitas. A useapi.net custa $15 por mês, com todas as APIs e 3 contas Google por assinatura.
• Os exemplos geram vídeos e imagens em lote a partir de um prompts.json, montam uma cena longa com extensões do Omni 1.2 e criam um vídeo UGC de produto do início ao fim.

Artigos em português: Como usar a API do Veo 3.1 com curl pela Google Flow API e Dá para usar a assinatura do Google AI por API?. Documentação completa: Google Flow API.

Về repo này

Ví dụ Node.js và Python cho Google Flow API, một API của bên thứ ba do useapi.net host, để tạo qua REST video Veo 3.1 Fast, Quality và Lite, video Omni 1.2 và 1.1 Flash có âm thanh, và ảnh Nano Banana Pro, 2.1 và 2 Lite.

• Chạy trên gói Google AI của bạn: Ultra $199,99, Ultra $99,99, Pro $19,99, Plus $4,99 mỗi tháng hoặc gói miễn phí (giá tháng 11/2026), đã hỗ trợ tên miền mới flow.google.com.
• Không cần dự án Google Cloud hay Gemini API key, không phải tự triển khai server hay duy trì trình duyệt.
• Tài khoản Google được kết nối bằng quy trình trình duyệt tự động: đăng nhập Google một lần, không cần chép cookie.
• reCAPTCHA được server tự giải: tài khoản đầu tiên được 300 lần miễn phí, sau đó bạn dùng dịch vụ giải captcha của riêng mình.
• Chi phí: trên Google AI Ultra $199,99, một video Veo 3.1 Fast dài 8 giây tốn khoảng $0,08 credit Flow, so với $0,80 trên API chính thức của Google (tháng 11/2026). Ảnh không tốn credit và dùng được trên tài khoản Google miễn phí. useapi.net giá $15 mỗi tháng, gồm mọi API và 3 tài khoản Google mỗi gói đăng ký.
• Các ví dụ tạo hàng loạt video và ảnh từ một file prompts.json, dựng một cảnh dài bằng các lần kéo dài Omni 1.2, và làm trọn một video UGC giới thiệu sản phẩm.

Bài viết tiếng Việt: Cách dùng Veo 3.1 API: tạo video bằng curl qua Google Flow API và Dùng gói Google AI qua API được không?. Tài liệu đầy đủ: Google Flow API.