MuleRouter CLI

Command-line interface for MuleRouter and MuleRun multimodal AI APIs. Generate images, videos, speech, and music from the terminal.

Features

• 40+ model endpoints across 6 providers (Alibaba, Google, KlingAI, Midjourney, MiniMax, OpenAI)
• Text-to-Image, Text-to-Video, Image-to-Video, Text-to-Speech, Text-to-Music, and more
• Async task management with status command for non-blocking workflows
• Automatic task polling with progress updates
• Local image file auto-conversion to base64
• Retry with exponential backoff on transient errors
• JSON output mode for scripting and AI agent integration

Installation

npm install -g mulerouter
# or
bun add -g mulerouter

Quick Start

1. Configure

Set your API key and site:

export MULEROUTER_API_KEY=your-api-key-here
export MULEROUTER_SITE=mulerouter   # or "mulerun"

Or create a .env file:

MULEROUTER_API_KEY=your-api-key-here
MULEROUTER_SITE=mulerouter

Get an API key at mulerouter.ai/app/api-keys.

2. List available models

mulerouter list
mulerouter list --provider google
mulerouter list --output-type video
mulerouter list --tag SOTA --json

3. Check model parameters

mulerouter params alibaba/wan2.6-t2v
mulerouter params google/nano-banana-2/edit --json

4. Generate content

Text-to-Video:

mulerouter run alibaba/wan2.6-t2v --prompt "A cat walking through a garden"

Text-to-Image:

mulerouter run google/nano-banana-2/generation --prompt "A serene mountain lake" --resolution 2K

Image-to-Video:

mulerouter run alibaba/wan2.6-i2v --prompt "Gentle zoom in" --image /path/to/photo.png

Text-to-Speech:

mulerouter run minimax/speech-2.8-turbo --prompt "Hello world" --voice-id "Charming_Lady"

Text-to-Music:

mulerouter run minimax/music-2.5 --prompt "upbeat pop" --lyrics-prompt "[verse]\nHello world\n[chorus]\nLa la la"

Commands

mulerouter list

List available model endpoints with optional filtering.

Option	Description
--provider <name>	Filter by provider (alibaba, google, klingai, midjourney, minimax, openai)
--output-type <type>	Filter by output type (image, video, audio)
--tag <tag>	Filter by tag (e.g., SOTA)
--site <site>	Filter by site availability (mulerouter, mulerun)
--providers	List providers only
--json	Output as JSON

mulerouter params <endpoint>

Show parameters for a model endpoint.

mulerouter params <provider>/<model>              # auto-resolves action
mulerouter params <provider>/<model>/<action>      # explicit action

Option	Description
--json	Output as JSON

mulerouter run <endpoint>

Run a model endpoint to generate content.

mulerouter run <provider>/<model> [--param value ...]
mulerouter run <provider>/<model>/<action> [--param value ...]

Option	Description
--api-key <key>	Override API key
--base-url <url>	Override base URL
--site <site>	Override site (mulerouter, mulerun)
--no-wait	Create task without waiting for completion
--poll-interval <s>	Polling interval in seconds (default: 20)
--max-wait <s>	Maximum wait time in seconds (default: 900)
--quiet	Suppress progress output
--json	Output as JSON
--extra <KEY=VALUE>	Pass extra parameters

Model-specific parameters are passed as --param-name value (use mulerouter params to see available options).

mulerouter status <api-path> <task-id>

Check the status of an async task. Use with run --no-wait for non-blocking workflows.

# Submit a task without waiting
mulerouter run alibaba/wan2.6-t2v --prompt "A cat" --no-wait --json

# Check status later
mulerouter status /vendors/alibaba/v1/wan2.6-t2v/generation <task-id>

# Poll until completion
mulerouter status /vendors/alibaba/v1/wan2.6-t2v/generation <task-id> --wait

Option	Description
--wait	Poll until task completes
--poll-interval <s>	Polling interval in seconds (default: 20)
--max-wait <s>	Maximum wait time in seconds (default: 900)
--json	Output as JSON

mulerouter config

Show current configuration and setup help.

Configuration

Environment Variables

Variable	Required	Description
MULEROUTER_API_KEY	Yes	API key for authentication
MULEROUTER_BASE_URL	No	Custom API base URL (takes priority over MULEROUTER_SITE)
MULEROUTER_SITE	No	Site: mulerouter or mulerun (default: mulerouter)

Priority: CLI arguments > environment variables > .env file

Base URL resolution: MULEROUTER_BASE_URL > MULEROUTER_SITE

• mulerouter → https://api.mulerouter.ai
• mulerun → https://api.mulerun.com

Supported Models

Alibaba

Wan 2.6, 2.5, 2.2, 2.1 series, Happy Horse 1.0 — text/image-to-video, text/image-to-image, VACE editing, keyframe interpolation.

Google

Nano Banana, Nano Banana 2, Nano Banana Pro (T2I + edit), Veo 3 (T2V + I2V).

KlingAI

Kling V3 and V3 Omni — T2V, I2V, Ref2V, V2V, video editing with multi-shot and sound.

Midjourney

Diffusion (T2I) and Video (T2V/I2V).

MiniMax

Speech 2.8 HD/Turbo (TTS), Music 2.0/2.5 (TTM) — mulerun only.

OpenAI

GPT Image 2 (T2I + edit) — mulerouter only.

Image Input

For image parameters (--image, --images, --first-frame, etc.), local file paths are preferred:

mulerouter run alibaba/wan2.6-i2v --image /tmp/photo.png --prompt "Zoom in"

Local files are validated for safety, converted to base64, and sent to the API. Supported formats: PNG, JPG, JPEG, GIF, BMP, WebP, TIFF, SVG, ICO, HEIC, HEIF, AVIF.

Development

This project uses bun workspaces.

bun install          # install all dependencies
bun run build        # build both packages (core first, then CLI)
bun run test         # run all tests
bun run typecheck    # type check both packages
bun run lint         # lint with Biome

License

MIT