Where to Find OpenWork Documentation: Official Guides, README, and Source Docs

OpenWork documentation is distributed across three primary locations: the official website for user guides, the repository README for installation and quick-start instructions, and the /docs folder for architectural specifications and deployment details.

OpenWork is an open-source AI workspace platform maintained by different-ai/openwork. Whether you are configuring the Electron desktop client, integrating the MCP server with an AI agent, or deploying the OpenWork Den control plane, you need to know exactly where to find the authoritative OpenWork documentation for each component.

Official Website Documentation

The primary user-facing resource lives at openworklabs.com/docs. This site contains the complete getting-started guide, UI walkthroughs, feature overviews, and API reference materials. It tracks the dev branch and remains the most up-to-date source for end-user instructions, including links to the OpenWork Den control-plane documentation.

Repository README Quick Reference

For developers cloning the repository, the README.md at the repository root provides a concise project overview, download badges, and immediate installation steps. It covers how to run the Electron-based desktop application locally using pnpm dev and how to connect AI agents such as Claude Code, Codex, or Cursor via the MCP endpoint. The README also includes the opencode.json configuration snippet for registering the OpenWork MCP server.

Deep-Dive Architecture Docs in /docs

The /docs directory in the different-ai/openwork repository houses internal design specifications, deployment manifests, and architectural blueprints. These markdown files explain the system's underlying mechanics rather than end-user workflows.

Memory Bank and Marketplace Design

Two critical architectural documents detail the data persistence and distribution models:

• docs/memory-bank-architecture.md – Describes the persistent storage architecture used by the Memory Bank component.
• docs/marketplace-capabilities-architecture.md – Outlines the design for the marketplace and the capability publishing pipeline, including runtime flow and data models.

Extension Manifests and MCP Profiles

For extending OpenWork with custom skills or managing developer profiles:

• docs/extensions-manifest-foundation.md – Defines the JSON schema for extension manifests that power plugins and skills, referenced by the /.warden/skills directory.
• docs/mcp-ui-control-profile.md – Explains profile creation (OPENWORK_DEV_PROFILE), keychain mocking for credential-free testing, and MCP UI interactions.

Key Source Files for Developers

When examining the implementation details alongside the documentation, reference these specific source files:

File	Purpose
README.md	Project overview, MCP registration examples, and local development commands.
AGENTS.md	Describes external agent integration workflows and MCP usage patterns.
packages/app/src/main.ts	Entry point for the Electron desktop client.
/.warden/skills	Directory containing skill definitions referenced by the extension manifest docs.

Quick Start Code Examples

Registering the OpenWork MCP with an AI Agent

Add the following configuration to your agent's settings (e.g., OpenCode):

{
  "mcp": {
    "openwork": {
      "type": "remote",
      "enabled": true,
      "url": "https://api.openworklabs.com/mcp/agent",
      "oauth": {}
    }
  }
}

Starting Local Development

Launch the desktop application with the shared development profile:

pnpm dev

Or specify a custom profile and debugging ports:

OPENWORK_DEV_PROFILE=my-feature \
OPENWORK_ELECTRON_REMOTE_DEBUG_PORT=0 \
PORT=0 pnpm dev

Adding the MCP to Codex CLI

codex mcp add openwork --url https://api.openworklabs.com/mcp/agent

Summary

• Primary documentation for users is hosted at openworklabs.com/docs, covering installation, tutorials, and API references.
• Quick-start guidance resides in the repository README.md, including pnpm dev instructions and MCP connection examples.
• Architectural specifications are located in the /docs folder, detailing Memory Bank design, marketplace capabilities, extension manifests, and MCP UI profiles.
• Implementation entry points include packages/app/src/main.ts for the Electron client and AGENTS.md for agent integration patterns.

Frequently Asked Questions

Where is the official OpenWork documentation website?

The official documentation lives at https://openworklabs.com/docs. This site contains the user guide, installation steps, UI walkthroughs, and up-to-date API references tracking the dev branch.

What is the difference between the README and the /docs folder?

The README.md provides a high-level overview, quick installation steps, and basic MCP configuration examples. The /docs folder contains deep architectural specifications, deployment guides, and internal design documents such as the Memory Bank and Marketplace Capability architectures.

How do I configure the OpenWork MCP for my AI agent?

Configure your agent (Claude Code, Codex, Cursor, or OpenCode) using the MCP endpoint https://api.openworklabs.com/mcp/agent. For Codex, run codex mcp add openwork --url https://api.openworklabs.com/mcp/agent. For OpenCode, add the JSON configuration object found in the README to your opencode.json settings file.

What file contains the architecture for the Memory Bank system?

The Memory Bank architecture is documented in docs/memory-bank-architecture.md within the repository. This file explains the persistent storage model and data structures used by the OpenWork desktop application and Den control plane.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}