poco-ai/poco-claw

A more beautiful and easier-to-use alternative to OpenClaw. It features a nicer Web UI, built-in IM support, a sandboxed runtime and channel-based team collaboration. Under the hood, it is powered by a Claude Code–based agent.

What it solves

Poco provides a collaborative, agentic workspace that moves beyond simple chatbots. It solves the problem of managing complex, long-term AI tasks by providing a secure sandbox for code execution, project-level organization, and a shared environment where multiple users and agents can collaborate via channels and threads.

How it works

The system operates as a "pocket coworker" that integrates several advanced capabilities:

  • Agentic Execution: It uses a secure isolated container (sandbox) to run commands, install dependencies, and modify files without risking the host system.
  • Collaboration Framework: It organizes work into servers and channels, allowing users to mention persistent agents and track their progress in an execution drawer.
  • Tool Integration: It supports the Model Context Protocol (MCP), custom Skills, and a built-in browser for autonomous web research.
  • Memory & Context: It utilizes mem0 for smart memory, allowing the agent to remember user preferences and project context over time.
  • Deployment: It can be self-hosted via Docker or integrated with IM platforms like Telegram, DingTalk, and Feishu.

Who it’s for

  • Developers and teams who need an AI agent capable of executing code and managing files within a secure environment.
  • Users looking for a more polished, feature-rich alternative to OpenClaw.
  • Teams requiring collaborative AI workflows where agents can participate in shared channels and produce reusable artifacts.

Highlights

  • Secure Sandbox: Isolated containers with optional local directory mounting for direct project file access.
  • Collaborative UI: Support for channels, DMs, reply threads, and an artifacts view for rendering various formats (HTML, PDF, Markdown, etc.).
  • Agentic Tooling: Native Claude Code-like experience with Plan Mode, slash commands, and GitHub repository integration.
  • Extensibility: Easy import of MCP tools and custom skills.
  • Smart Memory: Personalized interactions powered by mem0.

Related

  • Project
    agent-infra/sandbox

    A unified agent sandbox environment that combines a browser, shell, and IDE tools in one Docker container with a shared filesystem for AI agents.

  • Project
    Team-Commonly/commonly

    An open-source workspace for collaborating with AI agents from any runtime, providing them with persistent identity and shared project memory to eliminate redundant context-setting.

  • Project
    opensandbox-group/OpenSandbox

    A general-purpose sandbox platform for AI applications that provides secure, isolated environments for code execution, agent evaluation, and RL training.

  • Project
    Simpleyyt/ai-manus

    A general-purpose AI agent system that executes tasks in a secure Docker sandbox with capabilities for web browsing, code execution, and file management.

  • Dispatch
    Securing AI Agents with Isolated Docker Sandboxes

    An exploration of agent-sandbox, a tool designed to run AI coding agents within isolated Docker containers to prevent unauthorized system access and ensure execution safety.