souLLMate // v0.9 private beta

Access your subconscious via your data footprint.

Your wearable, calendar, screen time and chats with AI already track what you do. souLLMate weaves them into a single digital loom — so the patterns you sense but can't name become visible.

15+ sources 100% local-first 0 data sold
Life stream · last 9 months
1,372hours
1,367episodes
84ksignals
Stress Screen Social Flow
Work
Friends
Hobbies
Health
JanFebMarAprMayJunJulAugSep
The gap

You have the data. It's just scattered across fifteen apps.

Each app knows a sliver of you. None of them talk to each other. The patterns that shape your life — why you crashed in March, why you're restless on Sundays — stay invisible.

01Your wearableTracks sleep and recovery in isolationWhoop · Oura
02Your calendarKnows where your hours go but not whyGoogle · Apple
03Your screen timeSees attention but misses intentiOS · macOS
04Your AI conversationsHas your thinking, zero context on youChatGPT · Claude
05Your messages & photosRecords emotion without structureiMessage · Photos
The loom

One fabric. Every thread of your life, woven.

souLLMate reads each source on your machine, categorises every episode by state — Stress, Screen, Social, Flow — and surfaces the correlations across them. No dashboards to babysit. Just the patterns your subconscious already feels.

01 · Patterns

Cross-source correlations

Sleep dips after late Slack. Focus peaks on Tuesdays. Your loom highlights the relationships between sources — statistically grounded, never guessed.

02 · Query

Ask your own data

"When do I do my best creative work?" "Which weeks was I burning out?" souLLMate answers from your footprint, not the internet's.

why am I tired on Mondays
03 · Energy

See where you actually go

Your calendar says 40 hours of work. Your loom shows 12 hours of making, 9 of reacting, 19 of meetings that could have been a message.

How it works

Three steps. Runs locally on your Mac.

Step 01 · Connect

Plug in your sources

Apple Health, Google Calendar, screen time, ChatGPT, Claude, iMessage metadata, Gmail headers, photo metadata, aTracker, Uber history, Garmin. You pick what to include.

Step 02 · Weave

Your loom assembles itself

Episodes get categorised by intent. Correlations run locally. The longer you've got history, the denser the fabric.

Step 03 · Read

Surface what you already feel

Patterns, anomalies, honest answers to the questions you've been asking yourself. No coaching. No nudges. Just clarity.

Supported sources
Apple Health
Whoop
Oura
Ultrahuman
Garmin
Google Cal
Apple Cal
aTracker
Screen Time
ChatGPT
Claude
Gmail
iMessage
Photos
Uber
Trust is the product

Your subconscious shouldn't live on someone else's server.

Local-first. Period.

Raw data never leaves your Mac. The cloud mirror only pulls your data down to your device — never the other way around. Your footprint stays yours.

Open framework

Your local code is fully inspectable. You can audit every line that touches your data — nothing hidden, nothing proprietary between you and your own patterns.

Zero training

Your data is never used to train any model — ours or anyone else's. No exceptions buried in terms.

Revocable everything

Disconnect any source, delete your loom, export your history. You own the off switch.

Get started

Pick the door in.

Blueprint
Waitlist
The self-serve toolkit. Connect your own sources, run your own loom.
  • souLLMate desktop app (Mac)
  • Open framework · inspect every line
  • All 15+ data sources
  • Local-first data store
  • Community support
Join waitlist
Private project
$1,000 · one-time
We build your full observability system across multiple sessions. Custom to your life.
  • Everything in Assessment
  • Multi-session engagement
  • Custom loom categories
  • Ongoing advisory access
  • Priority support
  • Early access to new sources
Apply →
Questions

Things people ask.

01What does "access your subconscious via your data footprint" actually mean?+
Your behaviour leaves traces everywhere: how you sleep, who you message, what you ask AI, how long you spend looking at what. These traces encode patterns about you that you don't consciously track. souLLMate reads them and reflects them back — not as advice, but as clarity.
02How is this different from Whoop or Oura?+
A wearable sees one slice. souLLMate weaves every slice. Whoop tells you recovery dropped. souLLMate shows you the late calendar block, the 11pm screen time, and the anxious chat thread that preceded it.
03Where is my data stored?+
On your device. souLLMate runs locally on your Mac. No cloud sync of raw data. The cloud mirror only pulls your data down so it's accessible to you — raw signals never leave your machine.
04Do you train on my data?+
No. Not ours, not anyone else's. Your raw conversations, calendar entries and health records never leave your Mac in identifiable form.
05What do I need to get started?+
A Mac and at least two sources to connect — for example a wearable plus a calendar, or screen time plus AI chat history. The more threads you add, the richer the weave.
06How long until I see useful patterns?+
If you import months of existing history, patterns emerge on day one. With fresh data only, expect meaningful signal within three to six weeks.
07Is this medical or therapeutic advice?+
No. souLLMate surfaces correlations in your own data. It is not a diagnosis, not treatment, and not a substitute for clinical care.
Join

Your footprint is already there. Come read it.

We open the private beta in waves. Join now and we'll wire you up when your slot is next.

No spam · unsubscribe in one click