DEDUPED Hot AI repos, turned into use cases Get the Monday 5

How to recall what you finished and where you left off with Openhistory

Search a private timeline of yesterday's foreground work to pick up unfinished tasks.

Who it is forA knowledge worker who switched between editors, browsers, and trackers all day and cannot remember the state of a task the next morning.

The problemWork is spread across many apps, so by the end of the day it is hard to recall what moved forward and what is still open.

The outcomeYou get a searchable record of the day with hourly and daily recaps, and you can find the last state of a task in minutes.

What you need first

  • A Mac running macOS 14 or later
  • Accessibility permission for OpenHistory in System Settings (needed for richer context)
  • For building from source only: Node.js 22 and Xcode with Swift 6.1 or later
  • Optional: Apple Intelligence on macOS 26 or later for on-device summaries, or an API key for OpenAI, Anthropic, or Kimi

Steps

  1. Download OpenHistory for Mac from Openhistory.sh, or build it from source with git clone https://github.com/ztratar/openhistory.git.
  2. On first launch, accept the privacy notice and choose Apple On-Device, a cloud provider, or Summaries off.
  3. Open Settings, choose Grant access, and allow OpenHistory in macOS Accessibility.
  4. Work normally, then open the timeline and check the hourly and daily recaps for the day.
  5. Use the built-in history search on a task name to find where you stopped.

Prompt for your agent

Using the OpenHistory search in the app, find the most recent activity about <task name> and list the last state of that work in three bullet points. Verify each bullet against a timeline item that shows its date and time before you report it.

Caveat. Email and Messages capture are selected by default during setup, so clear those selections first if you do not want them recorded.

Use this in your AI agent

Ask your agent

Using the Deduped MCP, get the setup steps for ztratar/openhistory and "Recall what you finished and where you left off"

Add Deduped to Claude Code once

claude mcp add --transport http deduped https://dedupedai.com/mcp

Cursor, ChatGPT, others

Share: X LinkedIn Hacker News Email Copy link

Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer