ModelsLab CLI

The official command-line interface for ModelsLab — manage accounts, discover models, generate AI content (image/video/audio/3D/chat), handle billing, and interact with the full platform from the terminal.

Agent-first design — every command works for both humans (colored tables, prompts) and AI agents (--output json, --jq filtering, semantic exit codes).

Installation

npm

npm install -g modelslab-cli

Ships the prebuilt binary for your platform as an optional dependency — nothing is compiled and no install script runs.

PyPI

pip install modelslab-cli

Platform wheels; no build step and no Python dependencies.

_{Both register the modelslab command. The modelslab packages on npm and
PyPI are the SDKs, not this CLI — hence the -cli suffix.}

macOS (Homebrew)

brew install modelslab/tap/modelslab

Windows (Scoop)

scoop bucket add modelslab https://github.com/ModelsLab/scoop-bucket
scoop install modelslab

Linux (apt/deb)

# Download the latest .deb from GitHub Releases
curl -fsSL https://github.com/ModelsLab/modelslab-cli/releases/latest/download/modelslab_linux_amd64.deb -o modelslab.deb
sudo dpkg -i modelslab.deb

Direct Download (any platform)

curl -fsSL https://modelslab.sh/install.sh | sh

Go Install

go install github.com/ModelsLab/modelslab-cli/cmd/modelslab@latest

Manual Download

Download the latest binary for your platform from GitHub Releases.

Platform	Architecture	Download
macOS	Apple Silicon (M1/M2/M3)	`modelslab_darwin_arm64.tar.gz`
macOS	Intel	`modelslab_darwin_amd64.tar.gz`
Linux	x86_64	`modelslab_linux_amd64.tar.gz`
Linux	ARM64	`modelslab_linux_arm64.tar.gz`
Windows	x86_64	`modelslab_windows_amd64.zip`

Update

# Check whether a newer release is available
modelslab update --check
# Download, verify, and install the latest GitHub release
modelslab update

The CLI also checks for updates periodically during normal human-readable commands and prints a short notice when a newer release is available. It stays silent for --output json, --jq, shell completions, and MCP server mode. Disable the startup check for one command with --no-update-check, or persistently with:

modelslab config set updates.auto_check false

Quick Start

# New user? Sign up first
modelslab auth signup --name "Your Name" --email you@example.com --password "..." --confirm-password "..."# Login to your account in the browser
modelslab auth login --browser
# Or login with email/password
modelslab auth login --email you@example.com --password "..."# Check your profile
modelslab profile get
# Search models
modelslab models search --search "flux" --per-page 5
# Generate an image
modelslab generate image --prompt "sunset over mountains" --model sdxl
# Check wallet balance
modelslab wallet balance
# Get JSON output
modelslab profile get --output json
# Filter with jq
modelslab models search --search "flux" --output json --jq '.[].model_id'

Commands

modelslab auth Manage authentication (login, signup, logout, tokens)
modelslab profile Manage your profile (get, update, password, socials)
modelslab keys Manage API keys (list, create, get, update, delete)
modelslab models Discover models (search, detail, filters, tags, providers)
modelslab generate Generate AI content (image, video, audio, 3D, chat)
modelslab billing Manage billing (overview, payment methods, invoices)
modelslab wallet Manage wallet (balance, fund, transactions, coupons)
modelslab subscriptions Manage subscriptions (plans, create, pause, resume)
modelslab teams Manage teams (list, invite, update, remove)
modelslab usage View usage analytics (summary, products, history)
modelslab config Manage CLI configuration (set, get, profiles)
modelslab mcp MCP server mode (serve, tools)
modelslab docs Access API documentation (openapi, changelog)
modelslab completion Generate shell completions (bash, zsh, fish, powershell)
modelslab update Update the ModelsLab CLI

Authentication

The CLI uses two credential types:

Bearer token — for control plane commands (auth, profile, billing, etc.)
API key — for generation commands (image, video, audio, etc.)

Signup & Onboarding

Agents can programmatically create accounts and manage the full auth lifecycle:

# 1. Create account
modelslab auth signup --name "Your Name" --email you@example.com --password "..." --confirm-password "..."# 2. Verify email (token from verification email)
modelslab auth verify-email --token <verification-token># 3. Login in the browser (auto-stores bearer token + API key in OS keychain)
modelslab auth login --browser
# Or use email/password
modelslab auth login --email you@example.com --password "..."# 4. Check status
modelslab auth status
# Password recovery
modelslab auth forgot-password --email you@example.com
modelslab auth reset-password --token <reset-token> --password "..." --confirm-password "..."# Token management
modelslab auth tokens list
modelslab auth tokens create --name "ci-token"

Existing Users

# Browser login opens Chrome, asks you to grant CLI access, and stores both credentials
modelslab auth login --browser
# Email/password login also gets both token and API key
modelslab auth login --email you@example.com --password "..."# Or set API key manually
modelslab config set api_key "your-api-key"# Use environment variablesexport MODELSLAB_API_KEY="your-api-key"export MODELSLAB_TOKEN="your-bearer-token"

Model Selection

Every generation command accepts a --model flag to specify which AI model to use. With 50,000+ models available, discovering the right one is a key workflow.

Discovering Models

# Search by name
modelslab models search --search "flux"# Filter by feature category
modelslab models search --feature imagen # Image generation models
modelslab models search --feature video_fusion # Video generation models
modelslab models search --feature audio_gen # Audio/voice models
modelslab models search --feature llmaster # LLM/chat models
modelslab models search --feature threed # 3D generation models# Filter by base model architecture
modelslab models search --base-model sdxl
modelslab models search --base-model flux
# Get detailed info about a specific model
modelslab models detail --id flux
modelslab models detail --id cogvideox

Popular Models

Model ID	Category	Description
`flux`	Image	Flux Dev — fast, high-quality images
`midjourney`	Image	MidJourney-style artistic images
`sdxl`	Image	Stable Diffusion XL base
`seedance-t2v`	Video	Seedance text-to-video
`seedance-i2v`	Video	Seedance image-to-video
`cogvideox`	Video	CogVideoX video generation
`eleven_multilingual_v2`	Audio	ElevenLabs multilingual TTS
`eleven_sound_effect`	Audio	ElevenLabs sound effects
`scribe_v1`	Audio	ElevenLabs speech-to-text
`meta-llama-3-8B-instruct`	LLM	Meta Llama 3 8B chat
`deepseek-ai-DeepSeek-R1-Distill-Llama-70B`	LLM	DeepSeek R1 reasoning
`gemini-2.0-flash-001`	LLM	Google Gemini 2.0 Flash

Setting a Default Model

# Set default model for image generation
modelslab config set generation.default_model flux
# Override per-command with --model
modelslab generate image --prompt "sunset" --model midjourney

Browse all models: https://modelslab.com/models

Output Modes

# Human mode (default) — colored tables
modelslab models search --search "flux"# JSON mode
modelslab models search --search "flux" --output json
# JSON + jq filtering
modelslab models search --search "flux" --output json --jq '.[].model_id'

Generation

# Text to image (specify model with --model)
modelslab generate image --prompt "sunset over mountains" --model flux
# Image to image
modelslab generate image-to-image --prompt "oil painting style" --model midjourney --init-image https://...
# Text to video
modelslab generate video --prompt "ocean waves crashing" --model seedance-t2v
# Image to video
modelslab generate video --prompt "slowly zooming in" --model seedance-i2v --init-image https://...
# Text to speech
modelslab generate tts --text "Hello world" --language en
# Chat completion (OpenAI-compatible)
modelslab generate chat --message "Explain quantum computing" --model meta-llama-3-8B-instruct
# Music generation
modelslab generate music --prompt "upbeat electronic music"# Check generation status
modelslab generate fetch --id 12345 --type image
# Skip polling (async)
modelslab generate image --prompt "sunset" --model flux --no-wait

MCP Server Mode

Use the CLI as an MCP server for AI assistants:

# Start MCP server (stdio)
modelslab mcp serve
# List available tools
modelslab mcp tools

Claude Desktop / Claude Code config:

{
"mcpServers": {
"modelslab": {
"command": "modelslab",
"args": ["mcp", "serve"]
}
}
}

Billing & Payments

# View billing overview
modelslab billing overview
# Add funds to wallet
modelslab wallet fund --amount 25
# Add a payment method (card tokenized via Stripe)
modelslab billing add-payment-method --card-number 4242424242424242 --exp-month 12 --exp-year 2027 --cvc 123
# Subscribe to a plan
modelslab subscriptions plans
modelslab subscriptions create --plan-id 10 --payment-method pm_...
# Redeem a coupon
modelslab wallet redeem-coupon --code WELCOME50

Configuration

# Set config values
modelslab config set defaults.output json
modelslab config set generation.default_model sdxl
# View all config
modelslab config list
# Manage profiles
modelslab config profiles list
modelslab config profiles use work

Config file: ~/.config/modelslab/config.toml

Shell Completions

# Bashecho'source <(modelslab completion bash)'>>~/.bashrc
# Zshecho'eval "$(modelslab completion zsh)"'>>~/.zshrc
# Fish
modelslab completion fish >~/.config/fish/completions/modelslab.fish
# PowerShell
modelslab completion powershell | Out-String | Invoke-Expression

Exit Codes

Code	Meaning
`0`	Success
`1`	General error
`2`	Usage error
`3`	Authentication error
`4`	Rate limited
`5`	Not found
`6`	Payment/billing error
`7`	Generation timeout
`10`	Network error

Environment Variables

Variable	Purpose
`MODELSLAB_API_KEY`	API key for generation endpoints
`MODELSLAB_TOKEN`	Bearer token for control plane
`MODELSLAB_BASE_URL`	Override base URL
`MODELSLAB_PROFILE`	Active profile name
`MODELSLAB_OUTPUT`	Default output format
`NO_COLOR`	Disable colored output

Development

# Build
go build -o modelslab ./cmd/modelslab/
# Run tests
go test ./internal/... -v
go test ./tests/ -v # integration tests# Run with local server
./modelslab --base-url http://127.0.0.1:8888 auth login
# Release (with GoReleaser)
goreleaser release --snapshot --clean

License

MIT

Name		Name	Last commit message	Last commit date
.github/workflows		.github/workflows
cmd/modelslab		cmd/modelslab
internal		internal
packaging		packaging
tests		tests
.gitattributes		.gitattributes
.gitignore		.gitignore
.goreleaser.yml		.goreleaser.yml
AGENTS.md		AGENTS.md
CLAUDE.md		CLAUDE.md
README.md		README.md
go.mod		go.mod
go.sum		go.sum
install.sh		install.sh

Navigation Menu

Uh oh!

Repository files navigation

ModelsLab CLI

Installation

npm

PyPI

macOS (Homebrew)

Windows (Scoop)

Linux (apt/deb)

Direct Download (any platform)

Go Install

Manual Download

Update

Quick Start

Commands

Authentication

Signup & Onboarding

Existing Users

Model Selection

Discovering Models

Popular Models

Setting a Default Model

Output Modes

Generation

MCP Server Mode

Billing & Payments

Configuration

Shell Completions

Exit Codes

Environment Variables

Development

License

About

Releases

Packages

Contributors

Languages

Uh oh!

Latest commit

History

Folders and files

Repository files navigation

ModelsLab CLI

Installation

npm

PyPI

macOS (Homebrew)

Windows (Scoop)

Linux (apt/deb)

Direct Download (any platform)

Go Install

Manual Download

Update

Quick Start

Commands

Authentication

Signup & Onboarding

Existing Users

Model Selection

Discovering Models

Popular Models

Setting a Default Model

Output Modes

Generation

MCP Server Mode

Billing & Payments

Configuration

Shell Completions

Exit Codes

Environment Variables

Development

License

About

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages