Who's using what
Out of everyone signed in. Tap To do for the nudges that would help.
Checking…
Free plans: how full?
Each service gives a free amount. Amber means past 70%, orange past 90%. Tap a bar to see what it means.
Thoughts caught per day
Last 30 days.
Money
This month so far, the AI and speech used are worth —. You pay ₹0 today, because every service is on a free plan.
That's about — per thought caught, and — per person over the last 30 days.
The plan budgets up to — a month per device for this (the worst case in the October product review).
Waiting for your OK
Thought Catcher™ watches how people use it and suggests what would help. Nothing is sent or changed until you approve.
Weekly AI look
Once a week the AI reads everyone's usage numbers (never what they said) and suggests a few cards below, marked "Weekly AI look". You approve each one.Nothing waiting. Everyone's set up, and nothing needs fixing.
Send a message
Shows as a small card in their app next time they open it. With "also buzz", it's a phone notification too (only for people who turned notifications on, and never at night).
Do it by itself
Switch on the ones you trust. They then run every 5 minutes without asking (still at most one card per person every 3 days, never the same one twice in 2 weeks, never buzzing at night). Everything done shows in the log below.
What happened
Cards sent and whether people saw or used them, and everything done from this page.
| When | Card | To | Result |
|---|
Log
Bring people in
Make a join link and send it on WhatsApp. Whoever opens it taps "Continue with Google" and they're in: no setup, no waiting for you.
Or add someone's Gmail
Team and testers
Catcher devices
Recordings that came from a Thought Catcher™ device (not a phone), last 30 days. Set where each one sits; the brain uses it as context.
Who's using it
One row per person (all their phones together). Tap a heading to sort, or tap a person for a summary.
Checks
The things that should always be true.
What went wrong lately
Last 14 days, grouped. Most fix themselves: Thought Catcher™ tries again.
Nothing went wrong in the last 14 days.
Everything else, for when you want to dig in. Tap a section to open it.
AI use per day
"Tokens" are the small word-pieces AI counts in. One recording uses about 5,000. Counted from 7 Nov 2026.
Every AI and web call (7 days)
"Busy" = the service said "too many right now"; it's tried again and isn't a failure.
No calls in the last 7 days.
How well the brain understands
"Not sure" = it was less than 70% sure, or offered 2–3 meanings to pick from. Lower is better.
| Brain version | Since | Thoughts | How sure | Not sure | Removed by you |
|---|
What gets caught, and when
Kinds of thoughts
What happened to them
What Thought Catcher™ did or offered
Hour of day (darker = night)
Research and Notion
App requests and database
How often each part of the cloud was asked for something (last 30 days, from 7 Nov), and how big each table is.
Database tables
Limits and prices used
The free amounts the bars measure against, and the paid prices behind the ₹ numbers. Ask Claude to change one when a plan changes.
| Note |
|---|
How the numbers are counted
- Tracking
- Every call to Groq (AI), Sarvam (speech), Tavily (web search) and Notion is written down with what it used and how long it took. Counting began —; speech from before that was added from the recordings.
- Days
- Days are India time. Groq's daily allowances are measured over the last 24 hours.
- Money
- "Worth" is what the same calls would cost at the paid prices above. Sarvam's comes out of its free credit.
- Not measured
- Vercel (this page and the phone page) and Supabase's data transfer.