Smol Developer (smol-ai/developer) is an open-source, lightweight autonomous software development agent created by smol-ai. Designed as an early benchmark and minimal prototype for whole-codebase generation, it orchestrates LLM calls to translate natural language specifications directly into scaffolded repositories with multiple interconnected files. In practical usage, Smol Developer runs via a Python CLI interface or direct script execution. The workflow operates hierarchically: given a high-level product prompt, the agent first generates a shared technical specification and file structure list, then sequentially or concurrently generates the code for each planned file while referencing the shared context. It primarily targets OpenAI models via API keys configured in standard environment files. What distinguishes Smol Developer from heavyweight enterprise coding harnesses is its radical simplicity and transparency. Instead of maintaining complex multi-agent state machines or obscure cognitive loops, it provides a compact, readable codebase in Python that developers can easily read, modify, and embed directly inside custom internal workflows or domain-specific code generators. Smol Developer is distributed as open-source software under the MIT license and is free to use, requiring only the user's personal API credentials (such as an OpenAI API key) for compute. While highly influential in popularizing autonomous scaffolding agents, it remains a minimal reference implementation rather than a persistent background coding daemon.
Specification-driven generation — Translates a plain English prompt into a shared architectural design document before writing individual files. - Multi-file project scaffolding — Automatically creates directories, boilerplate, and source files mapped to the planned project structure. - Minimal Python footprint — Compact, auditable Python codebase that can be inspected and adapted into custom LLM pipelines. - Modular smol_dev library — Provides reusable Python modules alongside the standalone CLI runner. - BYOK (Bring Your Own Key) — Operates directly against LLM APIs using standard local environment variables without proprietary backend dependencies.
Use cases
Use Case: Rapid Green-Field Prototyping Developers can supply a markdown prompt describing a full prototype—such as a Pong game or utility API—and Smol Developer writes the directory structure, logic files, and configurations in a single run.
Use Case: Custom Coding Harness Foundation Engineers building specialized internal code generators can use the minimal smol_dev architecture as a starting boilerplate, customizing system prompts and context injection without managing complex agent frameworks.
Use Case: Exploring Multi-Step Generation Patterns Researchers and tool builders can test how standard LLMs handle whole-project cohesion by evaluating how shared specifications improve consistency across sequentially generated code files.
Who it is for
Smol Developer is best suited for engineers, researchers, and hobbyists who want a simple, transparent reference implementation to understand or fork for whole-repository code generation. It is not intended for teams looking for interactive IDE-integrated debugging agents, persistent background coding assistants, or complex refactoring workflows on existing enterprise codebases.
Install guide
Installation Clone the repository and install dependencies using Poetry or standard pip: bash git clone https://github.com/smol-ai/developer.git cd developer poetry install
First Run Configure your API credentials by copying the example environment configuration: bash cp .example.env .env Edit .env to include your OpenAI API key: bash OPENAI_API_KEY="sk-your-openai-key-here"
Example Session Run the agent via main.py by providing a prompt string describing your application: bash python main.py "build a simple cli todo application in python with sqlite persistence and unit tests" The agent will outline the necessary files, create the output directory, and stream the generated code files into your workspace.
Plandex — Plandex is an open-source terminal-based AI coding agent designed to handle complex, multi-file engineering tasks in large codebases. Created by…
OpenCode — OpenCode is an open-source AI coding agent developed by Anomaly that operates across terminal environments, desktop applications, and IDE extensions.…
OpenHands — OpenHands (formerly OpenDevin) is an open-source autonomous AI software development agent and orchestration platform. It provides an extensible runtime environment…
Subtask — Subtask is an open-source CLI agent harness and orchestration tool developed by zippoxer in Go. It is built to coordinate…
EchoCoding — EchoCoding is an open-source agent harness and audio-augmented developer tooling layer created by launsion-boop. Designed to run alongside AI coding…
pi-reflect — pi-reflect is an open-source agent harness and self-improvement extension for pi coding agents, developed by jo-inc. It inspects agent session…