Littlebird launches Mac app that builds searchable memory from your screen

In short: Littlebird has launched a Mac app that reads the text and layout of what you are actively viewing, then lets you search and ask questions about it later.

What happened

Littlebird is a new Mac app, with Windows support planned, that watches only your active window and captures what is on it every two seconds. It does this using macOS accessibility tools, the same kind of system features that help screen readers understand what is on a page (like labels, headings, buttons, and text).

Instead of taking screenshots or recording video, Littlebird reads the underlying text and structure of the app you are using. It ignores minimized apps, private browser windows, and sensitive fields like passwords by recognizing them in the accessibility “tree” (a map of what is on screen at a code level).

The app stores this activity as a private, searchable memory. You can ask it to find things you saw earlier, like a specific Slack link, or to summarize recent conversations or projects. It can also help draft emails or replies using that context, and it can generate daily briefings.

Littlebird says data is processed on your computer, but saved in encrypted form on AWS cloud servers. Users can pause collection, delete recent history like the last hour, or delete everything. The company says the data is not used to train outside AI systems, and it can use several AI models for answers, including Gemini, Claude, Llama, and GPT.

Why it matters

Many people lose time hunting for information they know they saw, but cannot quickly find again. Tools like this aim to turn your computer activity into a searchable diary (like having a “find” button for your workday), but they also raise privacy questions, so the opt outs, filters, and deletion controls will matter in real use.

Source: TechCrunch AI