OpenModels CLI

Command-line interface for the OpenModels registry — browse models, compare providers, and check telemetry from your terminal.

Installation

# Run without installing
npx @openmodels/cli

# Install globally
npm install -g @openmodels/cli

Requires Node.js ≥ 22.0.0.

Quick Start

# List all models
openmodels models list

# Get details about a specific model
openmodels models info claude-opus-4-6

# Compare providers for a model
openmodels compare gpt-4o

# Search models
openmodels search "code generation"

Commands

models list

List models with optional filters.

openmodels models list
openmodels models list --capability chat
openmodels models list --modality text --sort context_window
openmodels models list --page 2 --limit 20

Option	Description
--capability <cap>	Filter by capability (e.g., chat, code, reasoning)
--modality <mod>	Filter by modality (e.g., text, image, audio)
--sort <field>	Sort by: name, recency, context_window, providers
--page <n>	Page number
--limit <n>	Results per page

models info <model-id>

Display detailed information about a specific model including capabilities, modalities, context window, licensing, and provider availability with pricing.

openmodels models info claude-opus-4-6
openmodels models info gpt-4o --format json

providers list

List inference providers.

openmodels providers list
openmodels providers list --search anthropic
openmodels providers list --page 1 --limit 10

Option	Description
--search <term>	Filter providers by search term
--page <n>	Page number
--limit <n>	Results per page

compare <model-id>

Compare all providers offering a specific model side-by-side, showing pricing, rate limits, and regions.

openmodels compare gpt-4o
openmodels compare claude-opus-4-6 --sort rate_limit
openmodels compare deepseek-r1 --format yaml

Option	Description
--sort <field>	Sort by: price (default), rate_limit

search <query>

Full-text search across models by name or description.

openmodels search "code generation"
openmodels search "vision" --capability chat --limit 5

Option	Description
--capability <cap>	Additionally filter by capability
--limit <n>	Maximum number of results

telemetry <provider-id>

View provider health status and latency metrics.

openmodels telemetry anthropic
openmodels telemetry together-ai --period 7d
openmodels telemetry openai --format json

Option	Description
--period <period>	Time period: 1d, 7d, 30d

Output includes current status (up/down), uptime percentage, average time-to-first-token, average total response time, and p95 percentiles.

completions <shell>

Generate shell completion scripts.

# Bash
openmodels completions bash >> ~/.bashrc

# Zsh
openmodels completions zsh > ~/.zfunc/_openmodels

# Fish
openmodels completions fish > ~/.config/fish/completions/openmodels.fish

Output Formats

All commands support the --format flag to control output:

openmodels models list --format table   # Default — human-readable aligned columns
openmodels models list --format json    # Machine-readable JSON
openmodels models list --format yaml    # YAML output

Table format uses colored headers and status indicators for interactive use. JSON and YAML formats are designed for piping to other tools and CI/CD pipelines.

Configuration

Config File

Create ~/.openmodelsrc to store default settings:

{
  "api_url": "https://api.openmodels.run",
  "api_key": "om_...",
  "format": "table",
  "no_color": false
}

All fields are optional. If the file contains invalid JSON, the CLI prints a warning and continues with defaults.

Environment Variables

Variable	Description
OPENMODELS_API_URL	API base URL (default: https://api.openmodels.run)
OPENMODELS_API_KEY	API authentication key
NO_COLOR	Disable colored output (any value)

Precedence

Configuration is resolved with the following priority (highest to lowest):

1. CLI flags (--api-url, --api-key, --format, --no-color)
2. Environment variables (OPENMODELS_API_URL, OPENMODELS_API_KEY)
3. Config file (~/.openmodelsrc)
4. Defaults (api_url: https://api.openmodels.run, format: table)

Global Options

Flag	Description
--format <type>	Output format: table, json, yaml
--api-url <url>	API base URL
--api-key <key>	API authentication key
--no-color	Disable colored output
--verbose	Show detailed error information including stack traces
-h, --help	Show help
-V, --version	Show version

Error Handling

The CLI provides clear, actionable error messages:

• Resource not found — when a model or provider ID doesn't exist
• Rate limited — suggests waiting before retrying
• Network errors — suggests checking connectivity
• Unexpected errors — use --verbose for full details

All errors are written to stderr so piped output remains clean. The CLI exits with code 0 on success and non-zero on failure.

Development

cd cli

# Install dependencies
npm install

# Build
npm run build

# Run in development (watch mode)
npm run dev

# Run tests
npm test

# Type check
npm run type-check

# Lint
npm run lint

License

MIT