Your Life, Organized by Day

Everything you said, did, spent, and felt — automatically captured and organized into daily summaries. Scroll through your timeline. See when you were productive, what you spent, how you felt. Nothing to write, nothing to log. Sathi does it all in the background so you can look back anytime.

Four Layers, Zero Effort

4-Layer History Architecture💬Layer 1 — Raw InputEvery message stored with processed=false📋Layer 2 — Structured EventsActivityLog entries: task, expense, call, note, reminder, event📊Layer 3 — Daily AggregationDailyHistory bucket: counts, totals, emotional trend, top topics🧠Layer 4 — Background ExtractionEnd-of-day processing + diary generation + behavior metricsTime-of-Day View☀️Morning5 AM – 12 PM🌤️Afternoon12 PM – 5 PM🌙Evening5 PM – 9 PM🌑Night9 PM – 5 AMAggregated Per Day💬MessagesTasks💰Expenses📞CallsWhat You Get📄Daily SummaryMessage counts, task completion, spending totals😊Emotional TrendDominant mood of the day, auto-detected📔Auto DiaryAI-written diary from your day's data

History That Builds Itself

Date-by-Date Timeline

Scroll through your days like a calendar. Each date shows a collapsible card with everything that happened — messages, tasks created, money spent, calls made. Today auto-expands. Yesterday is one tap away.

Time-of-Day Grouping

Activities are organized into four natural periods: Morning (5 AM–12 PM), Afternoon (12–5 PM), Evening (5–9 PM), Night (9 PM–5 AM). See when you were most active, productive, or reflective.

Daily Aggregation

Message counts, task completion ratios, total spending, call logs — all aggregated into a single DailyHistory bucket per day. No need to dig through individual conversations.

Emotional Trend

Sathi detects the dominant emotion of each day from your conversations. See a mood indicator for every date — were you stressed? Happy? Calm? Patterns emerge over weeks.

Auto-Generated Diaries

At the end of the day, Sathi writes a diary entry from your data — tasks done, money spent, moods felt, topics discussed. You don't have to write anything. Your day is already documented.

Behavior Learning

Over a 7-day sliding window, the history engine computes behavior metrics — voice vs. text ratio, active hours, feature usage, task completion rate. These adapt your experience silently.

How It Works

1

Raw Input

Every message you send — voice or text — is stored locally in Realm with a processed=false flag. Nothing leaves your device.

2

Pattern Extraction

Local regex patterns detect tasks, expenses, calls, reminders, and events in your messages. No LLM needed for this step.

3

Daily Aggregation

All data for a date is counted, summed, and stored in a DailyHistory record. Total messages, tasks created vs. completed, spending, emotional trend, top topics.

4

End-of-Day Processing

At 11:59 PM (or when the app backgrounds), unprocessed messages are extracted, aggregated, and optionally turned into a diary entry using the AI.

Your Story Writes Itself

Download VivaLyn Sathi and let your history build in the background.