A terminal coding agent for real project work.

Peezy CLI

Peezy CLI connects your terminal to the Peezy Agent Gateway. It provides a focused coding-agent workflow with browser login, model discovery, usage tracking, repository-aware file mentions, and optional managed tools for image generation and web research.

Install

npm install -g @p0systems/peezy
peezy

The package installs the peezy binary.

Quick Start

Sign in with Peezy to get free starter usage through the hosted gateway:

peezy login
peezy auth status
peezy models
peezy "summarize this repo"

Prefer to use your own provider instead of Peezy login:

peezy providers
peezy config use <provider-id>
peezy config set apiKey <key>
peezy models

Run a single non-interactive prompt:

peezy --print "summarize @src/index.ts"

Peezy Go

Peezy login includes free starter usage through the hosted gateway. Peezy Go adds a larger monthly usage allowance, managed image generation, and stronger default model access. Run peezy upgrade or /upgrade inside the TUI to open the Peezy Go dashboard.

Configuration

The first run creates ~/.peezy/config.json with 0600 permissions when the filesystem supports it. Browser login stores a Peezy gateway key in that file. Do not commit it.

Environment variables can override config values:

• PEEZY_API_KEY
• PEEZY_BASE_URL
• PEEZY_MODEL
• PEEZY_PROVIDER
• PEEZY_WEB_URL
• OPENAI_API_KEY
• OPENROUTER_API_KEY
• AI_GATEWAY_API_KEY
• VERCEL_OIDC_TOKEN
• CUSTOM_OPENAI_API_KEY

PEEZY_API_KEY, PEEZY_BASE_URL, and PEEZY_MODEL only apply when the active provider is peezy, so they are not sent to OpenAI, OpenRouter, or custom providers.

Useful commands:

peezy config show
peezy config path
peezy config use openai
peezy config use vercel
peezy config set apiKey <key>
peezy providers
peezy credits
peezy models
peezy tools
peezy mcp
peezy upgrade

Built-in providers are peezy, openai, openrouter, vercel, local, and custom. Provider profiles use OpenAI-compatible APIs.

Provider Support

Peezy CLI supports the Peezy Agent Gateway and providers with OpenAI-compatible APIs. This build supports 106 unique provider profiles: 6 built-ins plus 100 additional runnable providers from the live Models.dev feed. Run peezy providers to list configured providers plus the live providers Peezy can actually connect to.

Connect a supported provider:

peezy providers
peezy config use <provider-id>
peezy config set apiKey <key>
peezy models
peezy config set model <model-id>

For one-key access to a large multi-provider model catalog, use Vercel AI Gateway:

export AI_GATEWAY_API_KEY=<key>
peezy config use vercel
peezy models
peezy config set model anthropic/claude-sonnet-4.6

Native-only provider adapters are not bundled, so providers without an OpenAI-compatible API are not shown by peezy providers. Use an OpenAI-compatible gateway such as Vercel AI Gateway or OpenRouter, or configure your own compatible endpoint with custom:

peezy config use custom
peezy config set baseUrl <openai-compatible-url>
peezy config set apiKey <key>
peezy models
peezy config set model <model-id>

TUI Commands

Inside the Peezy TUI:

• /model lists available models.
• /model <name> switches model and starts a fresh thread.
• /credits shows current usage.
• /tools shows available tools.
• /mcp shows connected MCP tools.
• /upgrade opens Peezy Go.
• @ opens file suggestions.

Headless prompts also expand file mentions:

peezy --print "explain @src/config.ts"

Development

npm install
npm run build
npm link
peezy --help

Run from source:

npm run dev -- --help
npm run dev -- --print "summarize this repo"

Use another provider without Peezy login:

OPENAI_API_KEY=<key> npm run dev -- --provider openai --print "summarize this repo"

Publishing

Before publishing:

npm run build
npm audit --audit-level=high
npm pack --dry-run

The npm package includes dist/, assets/, README.md, SECURITY.md, LICENSE, and package.json.

Security

Do not commit ~/.peezy/config.json, API keys, access tokens, or local .env files. See SECURITY.md for reporting guidance.

License

MIT