ClawWork  by clawwork-ai

Manage AI agent tasks with a parallel, multi-session desktop workspace

Created 1 week ago

New!

282 stars

Top 92.5% on SourcePulse

GitHubView on GitHub
Project Summary

ClawWork is a desktop workspace designed to enhance productivity for users of OpenClaw. It addresses the challenges of managing complex, multi-session AI workflows by providing an isolated, task-centric environment. This application targets power users and researchers who find standard chat interfaces insufficient for parallel task execution, offering improved visibility, organization, and control over AI agent interactions and outputs.

How It Works

ClawWork operates by connecting to an OpenClaw Gateway through its WebSocket API. Each distinct AI task is assigned its own isolated OpenClaw session, identified by a unique session key. This isolation prevents context switching issues and ensures that parallel jobs do not interfere with each other. The desktop application maintains a local SQLite index for task metadata, messages, and artifacts, enabling efficient full-text search and easy retrieval of past work, independent of chat logs.

Quick Start & Requirements

  • Installation: macOS users can install via Homebrew: brew tap clawwork-ai/clawwork && brew install --cask clawwork. Prebuilt macOS and Windows binaries are available on the Releases page.
  • Prerequisites: Requires a running OpenClaw Gateway. The default local endpoint is ws://127.0.0.1:18789.
  • macOS First Launch: If the application is blocked due to being unsigned, execute: sudo xattr -rd com.apple.quarantine "/Applications/ClawWork.app".
  • Setup: Basic setup involves launching ClawWork and configuring the OpenClaw Gateway connection.

Highlighted Details

  • Task-first workflow with parallel, isolated OpenClaw sessions.
  • Real-time streaming responses, inline tool call cards, and progress/artifact visibility.
  • Multi-gateway support, per-task agent/model switching, and approval prompts for sensitive actions.
  • Local artifact storage with full-text search across tasks, messages, and outputs.
  • Integrated local voice input via whisper.cpp.

Maintenance & Community

The project is described as "early and moving fast." Development guidelines are detailed in DEVELOPMENT.md. Community interaction primarily occurs through GitHub Issues and Pull Requests. A roadmap is available outlining future development.

Licensing & Compatibility

Licensed under the Apache 2.0 license. This permissive license generally allows for commercial use and integration into closed-source projects.

Limitations & Caveats

The project is in an early development stage. Linux packages are planned but not yet available. First-time launches on macOS may require a specific command to bypass quarantine restrictions.

Health Check
Last Commit

17 hours ago

Responsiveness

Inactive

Pull Requests (30d)
127
Issues (30d)
33
Star History
284 stars in the last 12 days

Explore Similar Projects

Feedback? Help us improve.