Summary

Zot is a lightweight, Go-based coding agent harness designed for developers and power users seeking a versatile tool for interacting with various Large Language Models. It offers a single static binary for easy deployment and integrates with over 30 LLM providers, including major cloud services, local models via Ollama, and subscription-based offerings. Zot provides essential tools for code manipulation and execution, flexible run modes, and extensive extensibility, enabling users to automate complex coding tasks efficiently.

How It Works

Built in Go and distributed as a single static binary, Zot employs a modular architecture to support an extensive list of LLM providers, from OpenAI and Anthropic to local Ollama instances and specialized APIs. Its core functionality revolves around four tools: read for file content, write for file creation, edit for in-place modifications, and bash for shell command execution. Zot operates in three primary modes: an interactive TUI for direct chat, a print mode for final output, and a json mode for structured RPC communication, facilitating integration into automated workflows.

Quick Start & Requirements

Installation is streamlined via one-liner scripts for macOS/Linux (curl -fsSL https://www.zot.sh/install.sh | bash) and Windows (iwr -useb https://www.zot.sh/install.ps1 | iex), or through go install github.com/patriceckhart/zot/cmd/zot@latest. Prebuilt binaries are available on the releases page. Authentication is handled via a /login command in the TUI, environment variables, or a local $ZOT_HOME/auth.json file, supporting both API keys and subscription-based OAuth flows. While Go is required for building from source, no other specific hardware or software prerequisites are mandated for running the binary.

Highlighted Details

• Broad LLM Integration: Supports over 30 LLM providers, including Anthropic, OpenAI, Gemini, Copilot, Mistral, Groq, and local models via Ollama.
• Core Toolset: Features read, write, edit, and bash tools for file manipulation and command execution within a sandboxed environment.
• Flexible Operation: Offers interactive TUI, print (stdout), and json (NDJSON for RPC) modes for diverse usage scenarios.
• Extensibility: Allows extensions in any language via JSON-RPC and custom instructions through SKILL.md files.
• Advanced Session Management: Includes features like session resuming, forking, branching (/session tree), and a "swarm" mode for concurrent sub-agent execution.
• Telegram Integration: A built-in Telegram bot bridge enables remote interaction with the agent.

Maintenance & Community

The project explicitly states "no community atm." Maintenance appears to be driven by the single author, with releases available on GitHub.

Licensing & Compatibility

Zot is released under the permissive MIT license, which generally permits commercial use, modification, and distribution without significant restrictions, making it suitable for integration into proprietary systems.

Limitations & Caveats

The use of provider-specific OAuth client IDs for subscription logins carries a risk of violating terms of service and potential revocation. Extensions are not installed by default and require explicit user action. The bash tool and file operations are sandboxed to the current working directory, but this sandbox is described as a guardrail rather than a strict security boundary.