Skip to content

medium_tutorial.md

Latest commit

 

History

History
80 lines (51 loc) · 2.74 KB

medium_tutorial.md

File metadata and controls

80 lines (51 loc) · 2.74 KB
Raw

Control Google Veo and Imagen from Your Terminal with gflow-cli

Google Flow provides access to Google's video model, Veo, and its image model, Imagen. Google exposes these models through a web interface. gflow-cli brings them to your terminal.

gflow-cli is an unofficial Python command-line interface. It uses a reverse-engineered private REST API to drive image and video generations. You can script, batch, and pipe your creative workflows.

Prerequisites

Before you begin, ensure you have:

  • Python 3.11 or newer.
  • The uv package manager.
  • A Google account with access to Google Flow.
  • A desktop environment to complete the one-time authentication.

Step 1: Install gflow-cli and Chromium

Install the CLI tool using uv:

uv tool install gflow-cli

Download the Chromium browser binary for Playwright:

uv tool run --from gflow-cli playwright install chromium

This download requires 150 MB of disk space.

Step 2: One-Time Authentication

Authenticate your Google account. Run the following command:

gflow auth login

This command opens a Chrome window. Google rejects any browser that advertises automation (navigator.webdriver) — not Playwright as such — and both the default auto and an explicit --browser chrome launch with the flags that keep it false. Log in to your Google account. Solve any CAPTCHA challenges.

Keep going until the Flow editor itself loads. gflow detects the completed sign-in and closes Chrome for you; closing the window yourself also works. The CLI saves your cookies in a user-local directory (for example, %LOCALAPPDATA%\gflow-cli\ on Windows or ~/Library/Application Support/gflow-cli/ on macOS).

Verify your session status:

gflow auth status

The output confirms your cookie status.

Step 3: Generate a Text-to-Image Still

Generate images using the image t2i command. Define your prompt and select an aspect ratio:

gflow image t2i "a serene mountain lake at dawn" --model nano-pro --aspect 16:9

Supported aspect ratios are 9:16, 16:9, 1:1, 4:3, and 3:4.

Model choices:

  • nano2: Fast generation with balanced quality. This is the default.
  • nano-pro: Higher quality.
  • image4: Photorealistic outputs.

The command saves the files to a date-partitioned folder inside your output directory. You can specify a custom output directory:

gflow image t2i "neon cyberpunk alley" --aspect 16:9 --out ./my-images

Step 4: Generate a Text-to-Video Clip

Generate a video using Google Veo with the video t2v command:

gflow video t2v "a steam locomotive moving through the snow at dusk" --aspect 16:9 --seed 4242

The --seed flag provides reproducibility. The command returns a standard .mp4 file in your output directory.