autoprogrammer

Autoprogrammer

Autoprogrammer is a software generator that turns a high-level software blueprint into a runnable Python project using AI agents.

Given a structured blueprint describing modules, data models, and behaviours, Autoprogrammer:

• Reads and validates the blueprint.
• Uses AI agent prompts to generate real Python packages, modules, and configuration files.
• Runs automated validation to catch syntax and basic logic issues.
• Outputs a ready‑to‑run Python project on disk.

All previous RPG / game‑style metaphors have been removed; this tool now focuses purely on software generation.

Features

• Blueprint‑driven generation (projects are defined declaratively).
• Multi‑agent AI workflow for code, tests, and docs.
• Automatic code validation (linting and basic checks).
• Deterministic output folder structure for easy deployment.

Requirements

• Python 3.10+ installed.
• An API key for the configured AI provider (e.g. Anthropic, OpenAI, etc.), exported as an environment variable.
• Git and a POSIX‑like shell (PowerShell or WSL on Windows work fine).

Installation

Clone the repository:

git clone https://github.com/smallville1979/autoprogrammer.git
cd autoprogrammer

Create and activate a virtual environment:

python -m venv .venv
source .venv/bin/activate    # Linux/macOS
# or
.\.venv\Scripts\activate     # Windows

Install dependencies:

pip install -r requirements.txt

Set your AI key (example for Anthropic):

export ANTHROPIC_API_KEY="your-key-here"   # Linux/macOS
# or
setx ANTHROPIC_API_KEY "your-key-here"     # Windows (new shell needed)

Usage

1. Write a blueprint file, for example blueprint.yaml:

project_name: example_app
description: Simple generated Python service
modules:
  - name: core
    description: Core domain logic
  - name: api
    description: HTTP API layer

1. Run Autoprogrammer against the blueprint:

python -m autoprogrammer \
  --blueprint blueprint.yaml \
  --output ./generated_project

1. Inspect the generated project:

cd generated_project
python -m pip install -r requirements.txt
python -m pytest          # if tests were generated
python main.py            # or the entrypoint created by the blueprint

The tool will:

• Parse the blueprint.
• Call the AI agents to generate Python modules, tests, and config.
• Validate the code.
• Write all files into generated_project/.

Development

To work on Autoprogrammer itself:

git clone https://github.com/smallville1979/autoprogrammer.git
cd autoprogrammer
python -m venv .venv
source .venv/bin/activate  # or .\.venv\Scripts\activate on Windows
pip install -r requirements-dev.txt

Run tests:

pytest

License

See LICENSE in this repository.