🎨 Best DeepSeek Harness Design Plugin. The open-source Claude Design alternative. 🖥️ Local-first desktop app. 🖼️ Your coding agent becomes the design engine: prototypes, landing pages, dashboards, slides, images & video — real files, HTML/PDF/PPTX/MP4 export. 🤖 Claude Code / Codex / Cursor / DeepSeek Harness / OpenCode & 20+ CLIs via BYOK.
Self-hosted AI agent app with MCP and skills
Coworker is a deployable AI agent product, not a build-time component. It gives you chat, scheduled workflows, MCP server management, a skills marketplace, Google Workspace access, and a desktop app for talking to the agent.
Builders who want a self-hosted agent app with chat, integrations, and workflow automation.
You can run an AI team member that answers questions, manages tasks, and connects to your tools from one app.
What it does
AI chat assistant
Answer questions, draft content, summarize documents, and brainstorm ideas in chat.
App builder
Create internal dashboards and tools that agents can maintain with git version control.
MCP UI and registry
Manage MCP servers in a visual UI and install servers from the built-in registry.
Skills marketplace
Install community-built skills from ClawHub and skills.sh.
A2A protocol
Let other agents discover and communicate with Coworker through Agent-to-Agent protocol.
Scheduled tasks
Run recurring workflows on a cron schedule through Inngest.
Google Workspace and WhatsApp
Work with email, calendar, docs, and WhatsApp messages through natural language.
Multi-provider model support
Switch between OpenAI, Anthropic, Gemini, NVIDIA, Groq, Kimi, and other OpenAI-compatible providers.
How to get it
- 1Download the desktop app for macOS and point it at your Railway server URL.
cp .env.example .env # add your API keys bun install bun run dev # http://localhost:4111
- 2Run
cd app && bun install && bun run dev
- 3The app connects to http://localhost:4111 by default. To connect to a remote server, go…
docker compose up
- 4The desktop app supports automatic updates via GitHub Releases. When a new release is…
src/mastra/ agents/ # Agent definitions tools/ # Reusable tools workflows/ # Scheduled tasks gog/ # Google Workspace integration whatsapp/ # WhatsApp bridge mcp/ # MCP server app/ # Electron desktop app
README
Coworker
Open-source AI agent built with Mastra. Acts as an AI team member that handles tasks, answers questions, and manages workflows via chat. Supports OpenAI, Anthropic, Google Gemini, NVIDIA, Groq, Kimi, and any OpenAI-compatible provider.
Features
- AI chat assistant — answer questions, draft content, summarize documents, and brainstorm ideas
- App builder — Lovable-like builder for creating internal dashboards and tools, maintained by agents with git version control
- MCP UI — visual interface for managing MCP servers and building agent-maintained internal dashboards
- MCP registry — discover and install MCP servers from the built-in registry
- Skills marketplace — install community-built skills from ClawHub and skills.sh
- A2A protocol — let other AI agents discover and communicate with Coworker via Agent-to-Agent protocol
- Scheduled tasks — run recurring AI workflows on a cron schedule via Inngest
- Google Workspace — manage emails, calendar, and docs through natural language
- WhatsApp bridge — interact with your AI agent via WhatsApp messages
- File workspace — upload, manage, and reference files in agent conversations
- API access — secure API key authentication for programmatic access and integrations
- Multi-provider AI — switch between OpenAI, Anthropic, Google, NVIDIA, Groq, Kimi, or any OpenAI-compatible endpoint
Screenshots
Stack
- Backend: Mastra agents + tools (TypeScript)
- Desktop: Electron app (React + Tailwind)
- Integrations: WhatsApp, Google Workspace (gog CLI), MCP
- Scheduling: Inngest
- Runtime: Bun
Deploy to Railway
One-click deploy of the Coworker server with Inngest, Postgres, Redis, and optional Tailscale private networking. After deploying, set your MODEL and API key (e.g. OPENAI_API_KEY) in the Railway service variables. See .env.example for all supported providers.
Download the desktop app for macOS and point it at your Railway server URL.
Self-Host
Backend
cp .env.example .env # add your API keys
bun install
bun run dev # http://localhost:4111
Desktop App
cd app && bun install && bun run dev
The app connects to http://localhost:4111 by default. To connect to a remote server, go to Settings > Advanced and update the Server URL.
Docker
docker compose up
Builds are automated via GitHub Actions and pushed to ghcr.io.
Auto-Updates
The desktop app supports automatic updates via GitHub Releases. When a new release is published, users are notified in Settings > Advanced where they can download and install the update.
Project Structure
src/mastra/
agents/ # Agent definitions
tools/ # Reusable tools
workflows/ # Scheduled tasks
gog/ # Google Workspace integration
whatsapp/ # WhatsApp bridge
mcp/ # MCP server
app/ # Electron desktop app
Files in the repo
- .agents
- .github
- app
- assets
- docker
- src
- .dockerignore
- .env.example
- .gitignore
- .mcp.json
- AGENTS.md
- bun.lock
- bunfig.toml
- CLAUDE.md
- CODE_OF_CONDUCT.md
- CONTRIBUTING.md
- coworker.pen
- docker-compose.yml
- Dockerfile
- index.ts
- LICENSE
- package.json
- README.md
- SECURITY.md
- tsconfig.json
Discussion (0)
Ask about usage, or say what you built with itSign in to join the discussion.
No comments yet. Be the first to say what this is good for.
More other
"Vibe-Trading: Your Personal Trading Agent"
OpenHuman is an open source personal AI for Mac, Windows and Linux — local-first memory, agent orchestration, and deep research.

Your Personal AI Assistant; easy to install, deploy on your own machine or on the cloud; supports multiple chat apps with easily extensible capabilities.
User-friendly AI Interface (Supports Ollama, OpenAI API, ...)
macOS video editor built for AI