Can I complete a task by name instead of ID?

Yes. complete_omnifocus_task accepts either a task ID or a task name. Just describe the task in plain language and the AI matches it against your live OmniFocus database, confirming before it marks anything done.

Will it complete the wrong task by accident?

No. The tool requires confirm=true, so your AI shows you the matched task and waits for confirmation before making any change. Nothing is checked off silently.

Does this work without OmniFocus open?

LMCP works with your local OmniFocus data on the Mac. Keeping OmniFocus running ensures the most reliable access, and your AI handles the rest through native macOS integration — no cloud sync required.

Complete an OmniFocus Task with AI on Mac

The complete_omnifocus_task tool lets your AI assistant mark an OmniFocus task as complete by task ID or by name. Instead of opening OmniFocus, hunting through projects, and ticking a checkbox, you just say what's done. Because finishing a task is a real change to your data, the tool requires confirm=true, so nothing is checked off until you say go.

A concrete example — you type into your AI client:

"Mark the OmniFocus task 'Send Q3 invoice to Acme' as complete."

The AI finds the matching task in your real OmniFocus database, confirms it, and marks it done. No app switching, no manual searching. Download LMCP to get started.

Which AI agents work?

LMCP exposes 150+ tools to the AI client of your choice. Desktop clients auto-configure over a local stdio connection: Claude Desktop, Cursor, VS Code (GitHub Copilot), Windsurf and Zed pick up LMCP automatically once installed. Web-based AIs connect through the LMCP Cloud Relay connector: ChatGPT, Claude.ai (web), Grok and Perplexity. Whichever you use, completing OmniFocus tasks works the same way — in plain language.

Automation

Completing a task rarely happens in isolation. complete_omnifocus_task chains naturally with the rest of the OmniFocus toolset. Your AI can list your tasks for today, identify which ones are finished, mark them complete in a batch, then add the follow-up tasks that came out of them. For example: "Show my flagged tasks, complete the three I already handled this morning, and create a new task to prep the client deck for Friday." The AI reads, reasons, and acts across all of it in one conversation.

Context

The AI isn't guessing or working from a stale export — it operates on your real, live OmniFocus database on your Mac. When you reference a task by name, it matches against the projects, contexts, and tags that actually exist in your setup right now. That means it can disambiguate "the invoice task" against your current task list, and it always confirms before changing anything, so a fuzzy name match never silently completes the wrong item.

Productivity

If you live in OmniFocus, the friction isn't deciding what's done — it's the clicks. Opening the app, navigating to the right perspective, finding the task, checking it off, and switching back to what you were doing adds up across a busy day. Letting your AI do it while you're already in a chat keeps you in flow. Clearing five completed items becomes one sentence instead of five context switches. Over a week, that's real minutes back and far less mental overhead.

Privacy & GDPR

Everything runs locally on your Mac. LMCP talks to OmniFocus directly through native macOS APIs — your tasks, projects, and notes never leave the machine and are never copied to any server. There are no API keys to manage and nothing is stored in the cloud. This is privacy by design: your data stays yours. Learn more about how LMCP is GDPR-compliant by architecture.