Zot is an open-source, terminal-based AI coding agent written in Go by Patrice Eckhart. Built to function directly within developer terminal environments, it provides an agentic coding interface capable of managing codebases, executing tasks, and orchestrating multi-turn conversational coding sessions. In daily use, Zot features an interactive terminal user interface (TUI) optimized for fast startup and session management. It supports session resumption with context usage restoration and transcript tail-rendering. Zot integrates across a wide variety of model providers, including Anthropic, OpenAI (including Responses API), Google Gemini, AWS Bedrock, Google Cloud Vertex AI, Azure OpenAI, and GitHub Copilot, alongside custom user-defined provider configurations. What sets Zot apart is its broad native provider flexibility combined with Model Context Protocol (MCP) support and experimental remote agent pairing capabilities (such as mobile phone pairing). By providing native Go binaries and cross-platform installation scripts, it avoids heavy runtime dependencies while remaining modular across multiple cloud and local model backends. Zot is licensed as open-source software and available publicly on GitHub. Because it operates on a bring-your-own-key model, users only pay standard API access fees directly to their configured model providers without platform markups.
Tags: agent harness, ai agent, cli, coding agent, developer tools, OpenSource