Async is an open-source developer tool that integrates AI coding, task management, and code review, aiming to streamline the workflow from GitHub issues to merged pull requests. It's designed for experienced developers working with mature codebases, offering an opinionated workflow that automates research, code execution in isolated environments, and code review through stacked diffs.

How It Works

Async utilizes a backend built with FastAPI and an asynchronous architecture. It leverages Google Cloud Run jobs for isolated task execution, employing Claude Code for implementation and other AI models for research and analysis. The workflow begins with importing GitHub issues, followed by an AI-driven research phase that analyzes the codebase and asks clarifying questions. Once confirmed, tasks are executed in isolated cloud environments, breaking down work into reviewable subtasks with separate commits. Code review is integrated within the app, allowing for comments and iterations on stacked diffs before merging.

Quick Start & Requirements

• Setup: Create and activate a virtual environment (uv venv .venv, source .venv/bin/activate), sync dependencies (uv sync), and install pre-commit hooks (pre-commit install).
• Local Development: Requires a .env file, a async-firebase.json config, and Google Cloud authentication (gcloud auth application-default login). Start the server with uvicorn src.server:app --port 8000.
• Prerequisites: Google Cloud Platform account (with Cloud Run enabled), Firebase project (with Firestore), GitHub App configuration, Stripe account, and API keys for Anthropic, OpenAI, and Google.

Highlighted Details

• Automates research, code execution, and review within a single application.
• Executes code changes in isolated Google Cloud Run jobs, preventing local environment impact.
• Breaks down complex tasks into reviewable subtasks with stacked diffs.
• Integrates directly with GitHub issues and pull requests.

Maintenance & Community

The project is open-source and accepts contributions via pull requests. Further community and support information can be found on their GitHub repository.

Licensing & Compatibility

The project is licensed under the MIT License, permitting commercial use and integration with closed-source projects.

Limitations & Caveats

The tool requires significant setup involving multiple cloud services and API integrations. While designed for mature codebases, its effectiveness may vary depending on the complexity and structure of the specific project.