A coding agent designed for efficiency and accessibility, Shelley offers a mobile-friendly, web-based interface for single-user coding assistance. It supports multi-conversation, multi-modal inputs (including screenshots and charts), and integrates with multiple AI models, aiming to provide a flexible and powerful tool for developers who need coding support on the go or across diverse AI capabilities.

How It Works

Shelley is built on a stack of Go for the backend, SQLite for data storage, and Typescript/React for the user interface. Its architecture revolves around Conversations, which contain Messages from users, models, tools, or a harness, all persisted in the SQLite database. Server-Sent Events (SSE) are used to efficiently update the UI, ensuring a dynamic and responsive experience. This design prioritizes bringing the agent's compute capabilities directly to the user's environment.

Quick Start & Requirements

Pre-built binaries for macOS and Linux are available via curl from the releases page, requiring chmod +x to make them executable. macOS users can also install via Homebrew (brew install --cask boldsoftware/tap/shelley). To build from source, clone the repository and ensure Go and Node.js are installed; then run make. Use make serve to start a local instance.

Highlighted Details

• Mobile-friendly, web-based interface for accessibility.
• Multi-modal support for inputs like screenshots, charts, and graphs.
• Multi-model integration to leverage diverse AI capabilities.
• Single-user architecture focused on local compute environments.

Maintenance & Community

New releases are automatically generated on every commit to the main branch. The provided README does not detail specific community channels (e.g., Discord, Slack), notable contributors, sponsorships, or a public roadmap.

Licensing & Compatibility

Shelley is licensed under the Apache License. Contributions require signing a Contributor License Agreement (CLA). The Apache License generally permits commercial use and integration into closed-source projects, though specific legal counsel is advised.

Limitations & Caveats

A critical limitation is that Shelley does not include built-in authorization or sandboxing features. Users are solely responsible for implementing their own security measures, particularly when deploying in shared or sensitive environments.