OpenSource03/harnss

Open-source, desktop client/UI build to harness Claude Code, Codex and any other Agent accepting Agent Client Protocol. Run multiple AI coding agents side by side with rich tool visualization, MCP integrations, built-in terminal, git, browser and just about anything else you may need.

What it solves

Harnss is a cross-platform desktop application designed to eliminate the friction of managing multiple AI coding agents. Instead of juggling separate terminals or losing context when switching between different AI tools, Harnss provides a single, unified interface for running and managing various agent engines.

How it works

The app acts as a management layer over existing agent engines. It supports built-in integration with Claude Code (via Anthropic SDK) and Codex, as well as any agent that follows the Agent Client Protocol (ACP). Users can point the app at a project folder on disk and run multiple agent sessions in parallel, each with its own state and history.

Who it’s for

Developers who use multiple AI coding assistants and want a centralized workspace that integrates their agent interactions with a terminal, browser, git management, and MCP (Model Context Protocol) servers.

Highlights

  • Multi-engine support: Run Claude Code, Codex, and ACP-compatible agents side-by-side.
  • Rich Visualization: Tool calls are rendered as interactive cards with word-level diffs and syntax highlighting rather than raw JSON.
  • Integrated Development Tools: Includes a built-in PTY terminal, embedded browser, and full git integration (staging, committing, pushing).
  • MCP Server Management: Connect MCP servers via stdio, SSE, or HTTP with automatic OAuth handling.
  • Agent Store: Browse and install agents from the community ACP registry directly within the app.
  • Permission Controls: Features a "Plan mode" for drafting changes and three levels of autonomy (Ask First, Accept Edits, Allow All).
  • Multimodal Input: Supports image attachments with a built-in annotation tool and voice input via macOS dictation or on-device Whisper models.

Related

  • Project
  • Project
  • Project
  • Project
  • Project