gobby

<!-- markdownlint-disable MD033 MD041 -->
<p align="center">
  <img src="logo.png" alt="Gobby" width="200" />
</p>

<h1 align="center">Gobby</h1>

<p align="center">
  <strong>Local-first daemon and workflow control plane for AI coding tools.</strong>
</p>

<p align="center">
  <a href="https://github.com/GobbyAI/gobby"><img src="built-with-gobby.svg" alt="Built with Gobby"></a>
  <a href="https://github.com/GobbyAI/gobby/blob/main/LICENSE.md"><img src="https://img.shields.io/badge/license-Apache%202.0-blue.svg" alt="License"></a>
  <a href="https://github.com/GobbyAI/gobby/stargazers"><img src="https://img.shields.io/github/stars/GobbyAI/gobby?style=flat" alt="Stars"></a>
  <a href="https://github.com/GobbyAI/gobby/issues"><img src="https://img.shields.io/github/issues/GobbyAI/gobby" alt="Issues"></a>
</p>

---

Gobby runs as a long-lived local daemon that unifies AI coding CLIs like Claude Code, Gemini CLI, and Codex, giving them shared sessions, memory, workflows, and guardrails instead of yet another one-off helper script. Because Claude Code natively supports OpenAI-compatible endpoints, local model providers like LM Studio and Ollama work out of the box — the same Gobby workflows run against both cloud and local models without changing your setup.

**Gobby is built with Gobby.** Most of this codebase was written by AI agents running through Gobby's own task system and workflows — over 10,000 tasks tracked and counting.

> Deterministic when you need it, autonomous when you do not — with hooks and workflows enforcing guardrails either way.

---

## Why Gobby exists

Modern AI coding tools are powerful but fragmented: each CLI has its own idea of sessions, context, and tasks, and none of them give you a single place to coordinate agents across tools. They also tend to burn tokens on redundant MCP metadata, verbose shell output, and repeated code snippets instead of the information that actually matters to the current change.

Gobby solves this by acting as the c