<2-3 sentences summarising what the project addresses and solves, the target audience, and the benefit.> GoGogot is a lightweight, self-hosted AI agent in Go, offering an open-source alternative to tools like OpenClaw. It enables users to deploy a personal AI assistant on their server, capable of executing shell commands, managing files, browsing the web, and maintaining memory, powered by LLMs like Claude or MiniMax. Its key benefit is a highly efficient, self-contained, and controllable AI agent experience with minimal resource overhead.

How It Works

Engineered as a single, ~15MB Go binary with minimal dependencies, GoGogot prioritizes simplicity and efficiency. It operates on an LLM-first philosophy, where the LLM dictates actions. The core loop processes input, queries an LLM for the next step (potentially tool execution), runs the selected tool (bash, file ops, web requests), appends results, and iterates. Behavior modification relies on prompt engineering, not complex code or plugins.

Quick Start & Requirements

• Primary install/run: Clone repo, configure .env (API keys for Anthropic/OpenRouter, TELEGRAM_BOT_TOKEN, TELEGRAM_OWNER_ID), then docker compose -f deploy/docker-compose.yml up -d.
• Prerequisites: Mandatory Telegram bot token/owner ID. API key for Anthropic Claude or OpenRouter provider (e.g., MiniMax). Docker image includes a full Ubuntu environment.
• Resource footprint: Idles at ~10MB RAM, active use ~30MB RAM.

Highlighted Details

• Extreme Efficiency: Single ~15MB binary, ~10MB idle RAM, ideal for low-cost VPS.
• Extensible Architecture: LLM providers and chat transports are Go interfaces, simplifying new integrations (e.g., Discord, Slack).
• Flexible Model Choice: Supports Claude and MiniMax (via OpenRouter) for cost/performance trade-offs via environment variables.
• Integrated Toolset: 17 built-in tools for system access (bash, file I/O), web operations (search, fetch), memory, and scheduling.

Maintenance & Community

No specific details on contributors, sponsorships, community channels (Discord/Slack), or roadmap were found in the provided README.

Licensing & Compatibility

Licensed under the MIT license, which is permissive for commercial use and integration into closed-source projects.

Limitations & Caveats

Agent behavior and new functionalities are primarily driven by prompt engineering, requiring prompt design expertise. Capabilities are tied to the chosen LLM provider's API and associated costs.