Skip to content

AI · Capture and recall

BrainSync AI remember in your own words

BrainSync is live in production. It is a second brain for people who think in fragments, not files. Capture a thought by text or voice, turn a spoken phrase into a reminder, and find it later by asking — not by remembering a folder path.

KotlinAICloud6 weeksIn production

The problem

Notes, alarms, and search usually live in three apps. A thought dies in a camera roll or a locked note. Retrieval had to feel like asking a person, not filing a document. No named client is listed.

What the product had to do

  • Make capture cheap: a plus button, voice, tags, and priority on a memories hub.
  • Create reminders from everyday phrases such as Call mom tomorrow 9am or Pay rent Friday noon — not from a calendar form.
  • Search in the user’s own words (doctor appointment, work meeting), including voice input and suggested queries.
  • Keep memories, reminders, and search in one app, with Settings as a fourth tab.

Challenge

Notes, alarms, and search usually live in three apps. A thought dies in a camera roll or a locked note. Retrieval had to feel like asking a person, not filing a document.

RR Tech's solution

Four tabs in a production app: Memories, Search, Reminders, Settings. Capture is cheap — a plus button, voice, tags, and priority. Reminders parse natural language so scheduling does not need a form. Search is built as a question box with voice and suggested queries, then filters across titles, tags, and notes.

Architecture and technology

Published stack: Kotlin, AI, Cloud. RR Tech’s Android work is Kotlin-first; this case study does not separately name iOS or a store.

Reminders parse natural language. Search is a question box, not a folder tree. That is product NLP as described — not a named foundation-model brand.

Cloud is listed; hosting provider and model vendor are not. We do not invent them.

Capture paths documented: text, voice, tags, priority.

Four tabs in production

Memories

Tagged notes, voice and search shortcuts, add-memory control. Outcomes also mention a timeline and connections. The public write-up does not specify how connections are computed.

Reminders

Empty state in the published screenshot invites phrases like Call mom tomorrow 9am, which become reminders automatically.

Search

Ask in your own words, voice input, suggested queries such as doctor appointment. Filters across titles, tags, and notes.

Settings

Shipped as the fourth tab. Extra settings behaviour is not documented on this site.

How it was built

The published timeline is 6 weeks, status in production. The documented approach is four tabs and cheap capture, then NL reminders and plain-language search. Studio phases remain Discover, Design, Build, and Launch. No sprint diary is published.

Results

Speak a thought once — get reminded, or find it by asking

  • Live memories hub with tags, voice capture, timeline, and connections
  • Reminders created from everyday phrases, not calendar forms
  • Plain-language search across everything you have saved

Status: in production. Outcomes are the memories hub, NL reminders, and plain-language search. No user counts, accuracy metrics, or revenue figures are published.

Like What You See? Let's Build Yours