Flux CLI

A command-line tool for AI image generation and editing using Flux through the AceDataCloud platform.

Features

• Text-to-Image Generation — Generate images from text prompts with multiple Flux models
• Image Editing — Edit existing images using text descriptions (kontext models)
• Task Management — Query status, batch check, and poll for completion
• Multiple Models — flux-dev, flux-pro, flux-kontext-pro, flux-kontext-max, flux-2-flex, flux-2-pro, flux-2-max, flux-2-klein
• Rich Output — Beautiful terminal formatting with --json for scripting

Installation

pip install flux-pro-cli

Quick Start

# Set your API token
export ACEDATACLOUD_API_TOKEN=your_token_here

# Generate an image
flux generate "A majestic mountain landscape at golden hour, photorealistic"

# Generate with a specific model
flux generate "Cyberpunk city with neon lights" -m flux-2-pro

# Generate with aspect ratio (ultra/kontext models)
flux generate "Minimalist logo of a phoenix" -m flux-2-max -s 16:9

# Generate multiple images
flux generate "A red car" --count 2

# Edit an existing image
flux edit "Add sunglasses to the person" --image-url https://example.com/photo.jpg

# Edit with max context model
flux edit "Change background to sunset beach" --image-url https://example.com/img.png -m flux-kontext-max

# Check task status
flux task abc123-def456

# Wait for task completion
flux wait abc123 --interval 5 --timeout 300

# Batch check multiple tasks
flux tasks task-1 task-2 task-3

Commands

Models

JSON Output

All commands support --json for machine-readable output:

flux generate "A red car" --json | jq '.task_id'
flux task abc123 --json | jq '.data[0].image_url'

Configuration

You can also use a .env file or pass --token directly.

Docker

docker compose run flux-cli generate "A beautiful sunset"

Development

# Install with dev dependencies
pip install -e ".[all]"

# Run tests
pytest

# Run linter
ruff check .
ruff format --check .

License

MIT License - see LICENSE for details.

Links

• AceDataCloud Platform
• API Documentation
• Flux by Black Forest Labs

FLUX 3 video

Save a JSON request such as {"action":"generate","mode":"t2v","prompt":"Waves at sunset","duration":5,"generate_audio":false} in request.json, then run flux video --request-file request.json. The request is validated before submission. Use i2v with keyframes, v2v with start_video, or draft_enhance with an owned platform draft_task_id. Draft availability is temporary.

Use flux video-edit --request-file edit.json with video and prompt, or flux video-upscale --request-file upscale.json with input_video, optional upscale_factor (1.5–3) and creativity (0/1). Upscale charges depend on actual output MP-seconds and frame rate. Commands output JSON and default to asynchronous submission; use flux task/flux wait to retrieve final video delivery. Set async=false in the request to wait synchronously.

All video operations use POST /flux/videos: action=generate (default), action=edit, or action=upscale. Generation additionally uses mode=t2v/i2v/v2v/draft_enhance. Each operation accepts only its own fields. The CLI utility commands and MCP utility tools set their matching action automatically.

The unified flux video --request-file request.json command also accepts editing and upscaling JSON when action is set explicitly. The video-edit and video-upscale commands remain convenient ways to submit those actions to the same endpoint.