Choose theme
AI Resources
OpenWork
OpenWork, from Different AI, is an OpenCode-powered workspace for AI agents working with files and connected apps. Its desktop application is offered for macOS, Windows and Linux, with model connections, skills and MCP tools.
You can use a local workspace or connect to an organization's managed setup. The organization's skills and connections can also be exposed through OpenWork's MCP Gateway to supported agent clients. Use this as a first read, not a recommendation. Open the original project before trusting details like terms, limits, privacy, cost, setup, or safety.
What it is
A workspace for agent tasks
Choose a working folder and model connection, then ask the agent to work with the material and tools available there. OpenWork supplies the interface around the agent, its instructions and connected capabilities.
Why it stands out
Local skills and managed sharing are separate paths
Local skills are folders containing instructions. Managed sharing adds an organization library, selected recipients and updates. That separates trying a method in your own workspace from maintaining it for a team.
Availability
Desktop releases and a public implementation
The repository provides desktop releases, application and server code, and executable tests. It names MIT and the OpenWork EE License; see the main repository for current terms. Organization services have separate setup and plan requirements.
Why it matters
What makes it useful
A work task may involve a folder of reference files, a draft and information from a connected service. OpenWork brings those inputs into an agent workspace. Its documented connected-file routes include bringing files from Google Drive, OneDrive or SharePoint into the workspace and saving edits back, with approval for transfers. Those routes depend on the service connection and its permissions; a local folder alone does not provide them.
What to know
Where it fits
The desktop app is the direct task interface. The MCP Gateway is another route for an organization's assigned capabilities: its tools let an agent search for a capability, execute it, or read an available skill. This route requires signing in and choosing an OpenWork organization. It is separate from adding a local MCP tool directly to one desktop workspace.
Notable points
What stands out
Sharing a connector and sharing an account are different choices. The connection guide offers per-person sign-in, where actions use each member's service account, or one account or API key for everyone given access. In the second mode, actions use that shared identity. Giving people the same connector therefore does not tell you whose account performs the work; the sign-in mode and that account's service permissions determine the access behind it.
Before using
What to review
The local desktop route does not require OpenWork Cloud sign-in, but the chosen model or service can still require an account, API key, local model setup or paid usage. A local workspace does not make a hosted model request local.
Managed sharing with access control and automatic updates uses OpenWork Cloud or a self-hosted deployment and the applicable team plan. Local skill folders can instead be copied manually; that does not carry managed access control or updates.
The MCP setup guide describes OpenCode as verified end to end and the other listed clients as setup-only. Treat those as the project's compatibility boundaries, rather than assuming every client path behaves alike.
Managed connectors use remote MCP URLs. Local stdio servers need setup on each machine.
Reader fit
Who may find it relevant
People working with files through an agent can start with the desktop application and a small folder task. Teams maintaining common instructions or service connections are the clearest audience for the managed library. Technical maintainers have a larger job when adding custom MCP servers or self-hosting the organization service; those are separate choices from simply installing the desktop app.
Editorial note
Why LifeHubber lists it
LifeHubber lists OpenWork for its managed skill-update path. The sharing guide says people who already have a shared skill receive its revised version on their next task, while a separate copy can be used to try a change first. For a team maintaining a recurring reporting method, this connects revising the instructions with getting that revision into later work. A correction can reach the next task through the assigned library, while an experiment stays in its own copy.
Source links
Source materials
Reader note
Before relying on this entry
LifeHubber lists entries to help readers inspect AI projects, not to endorse them or prove they are safe, suitable, accurate, maintained, or right for a specific use. We do not verify every entry in depth. Before relying on anything listed, review the original materials, terms, privacy practices, limits, and risks that matter for your situation.
More in AI Agents
Keep browsing this category
Explore more AI agent projects.
Paperclip
paperclipai/paperclip
A self-hosted server and dashboard for coordinating agent teams through companies, goals, roles, issues, heartbeats, budgets, approvals, and persistent activity records.
pixelfx
patch-gg/pixelfx
A code-driven pixel-art effects library with an optional Claude Code authoring skill, configurable palettes, sprite-sheet exports and timing data for character attacks.
Dash
agno-agi/dash
An Agno public data-agent template for asking questions over company data, with read-only SQL analysis, agent-created views in a separate schema, six context layers, a learning loop, AgentOS, Slack, Docker, Railway, and eval paths.
For project maintainers
Listed here? You can use the badge.
If you maintain a project with a current LifeHubber listing, you may add the optional “Listed on LifeHubber AI Resources” badge to its README, docs, or website. No introduction or permission request is needed.
