code-puppy

What it is and what it does

Code Puppy is a terminal-based AI code agent that acts as an alternative to commercial code editors. It connects to multiple LLM providers (OpenAI, Anthropic, Google, and others via models.dev) and uses tool calling to read files, execute shell commands, and generate code interactively. The agent understands programming tasks, generates code, and explains its reasoning in a conversational interface.

You interact with Code Puppy through a command-line REPL with slash commands (e.g., `/add_model` to browse and configure new LLM providers, `/meta-auth` for Meta OAuth, `/dbos` for optional durable execution). It supports custom commands via markdown files in project directories and can optionally checkpoint agent state to a database for crash recovery. The package is designed to be lightweight and scriptable, with low install friction via pure-Python distribution.

Use it for:

• Generate and refactor code interactively by describing tasks to an AI agent with access to your project files and shell.
• Experiment with multiple LLM providers and models without switching tools, using the integrated models.dev browser.
• Run long-running code-generation tasks with automatic crash recovery by enabling optional DBOS durable execution.
• Extend Code Puppy with custom slash commands by dropping markdown files into project directories.
• Use as a scriptable alternative to GUI-based code editors for CI/CD pipelines or headless environments.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Code Puppy is an AI-powered code generation agent that understands programming tasks, generates code, and explains its reasoning by integrating with multiple LLM providers and executing code through shell and file tools.

Yes, if you want a lightweight, open-source AI code agent with multi-provider LLM support and low install friction. The active maintenance, permissive MIT license, and broad provider integration make it a viable alternative to commercial tools. Install with caution on Android (Termux) due to compilation time and missing browser tools. Verify that your preferred LLM provider is supported and that you have valid API credentials before committing to it as your primary coding assistant.

Install

code-puppy on PyPI

pip install code-puppy

uv add code-puppy

poetry add code-puppy

Installing code-puppy

Before you install

Low install friction with a pure-Python wheel distribution. Active maintenance as of 2026-08-14 with 787 repository stars. Requires Python 3.11–3.14 and 23 runtime dependencies including anthropic, openai, boto3, and playwright, which may add setup time on first install.

License in practice

MIT license permits commercial and private use with minimal restrictions. You may use, modify, and distribute this package freely provided you retain the license notice.

Quickstart

# Install via UV (recommended)
uvx code-puppy -i

# Or via pip
pip install code-puppy

# Run interactively
code-puppy

# Then configure an API key (e.g., OpenAI):
# export OPENAI_API_KEY="sk-..."
# and use /add_model to browse and add models from providers

Requires Python 3.11 or later. First run may compile native dependencies (especially on Android/Termux). Playwright-backed browser tools are not available on Android. API keys for at least one LLM provider (OpenAI, Anthropic, etc.) are required to generate code.

Verify before relying

• How many providers are actually available through models.dev integration and which are thoroughly tested versus experimental.
• Performance and stability when running multiple concurrent agent instances.
• Whether the optional DBOS durable-execution plugin is production-ready and what database backends are recommended.

Package facts

Evidence: code_puppy-0.0.718-py3-none-any.whl

Tags

More Code Generators packages

Further reading