Home  /  Platform  /  Records

Platform · The atomic unit

If it isn't a typed record, it isn't in the ledger.

Records are the atom Lossless seals into the Provenance Ledger. Each one is typed, schema-versioned, lifecycle-linked, bi-temporal, and cryptographically signed. Open one in the browser and every field has a citation; no field is a blob. Re-fetch the source ten years from now and the seal still verifies.

See it work in Workspaces Read the next pillar → AI Pipeline
A slice of the real browser

Eight rows. Five record types. One viewer.

Open any cell and you reach the source: the Tesla service appointment, the Costco order line, the BofA statement PDF. No row is a summary; every row is a typed record.

Type Title Entity Date Sources Amount
Vehicle · service 12V battery replaced under warranty Tesla Service · Fremont · Apr 12, 2026 Model Y 2026-04-12 3 $0.00
Finance · charge Supercharging · Stockton 23 kWh · 41 minutes · session #4421 Model Y 2026-04-10 2 -$184.22
Email · receipt Costco · 4× tires Order #88421 · 60,000 mi warranty Model Y 2026-04-06 3 -$401.27
Voice "Just got new tires on the Y…" Auto-linked to Costco $401.27 charge Model Y 2026-04-06 2 linked
Statement · pdf BofA HELOC · March 14 line items · principal/interest split parsed 14 Pine Ave 2026-03-31 1 14 lines
Vehicle · tolls FasTrak · 6 trips Bay Bridge 4 attributed to 14 Pine Ave visit Model Y 2026-04-04 2 -$36.00
Recall Front camera firmware NHTSA · auto-watched · closed Apr 12 Model Y 2026-04-01 1 closed
Insurance · doc GEICO · annual renewal Coverage updated · premium $1,420/yr Model Y 2026-03-28 2 -$1,420

01 — A real records browser slice. Type, entity, date, sources, amount — all queryable, all signed, all addressable.

Domains, sub-domains, record types

Eight domains. Hundreds of record types. One viewer.

Records are organized as domain → sub-domain → record type. The "type" is the leaf — the most specific shape (vehicle.service.warranty, finance.statement.heloc). Each type has its own schema; all of them open in the same drawer.

Vehicle

Service · Charging · Tolls · Registration

Leaf types: service.warranty · service.paid · charging.session · toll.fastrak · recall.open · registration · smog. Linked to a specific car.

Finance

Transactions · Statements · Subscriptions

Leaf types: transaction.charge · transaction.transfer · statement.heloc · statement.brokerage · 1099-div · 1099-b · subscription.

Email

Receipts · Confirmations · Threads

Leaf types: email.receipt · email.confirmation · email.statement · thread · attachment. Auto-extracted with structured fields.

Property

Maintenance · Bookings · Leases · Tax

Leaf types: property.maintenance · property.booking · property.lease · property.expense · property.tax · payout.

Calendar & Travel

Events · Trips · Flights · Hotels

Leaf types: calendar.event · trip · flight · hotel · car-rental · trip.itinerary · travel.expense.

Documents

PDFs · Quotes · Extracted Entities

Leaf types: document.lease · document.permit · document.invoice · document.warranty · quote · extracted-entity. LLM-classified.

Power features

It's a browser — so it has the controls of one.

  • Filter by type, entity, date range, amount, source
  • Group by entity, vendor, category, month
  • Sort by date, amount, recency-of-change
  • Save any view as a named filter ("Vehicle YTD," "Pine Ave maintenance")
  • Export to CSV, JSON, or a printable PDF binder — every row carries its seal
  • Open any record to see the full source — original PDF, original email, original bank entry
  • Edit a typed field — Lossless tracks the version, never silently overwrites (bi-temporal)
  • Bulk tag, classify, or attach a vendor to many records at once

"I gave my CPA a saved view called 'Vehicle 2025.' She said it was the cleanest tax handoff she'd ever received."

— Beta user · two-vehicle household

One viewer for every record you'll ever own.

The browser is the surface. The discipline — typed, schema-versioned, sealed, bi-temporal — is underneath. Pick a workspace and see it doing real work.

See it work in Workspaces Read the next pillar → AI Pipeline
Continue the architecture tour

You've seen the brick. Now see the kiln.

Next pillar: how a raw inbox, a Tesla telemetry stream, or a 14-line PDF statement becomes one of these typed, signed records.

Next pillar → AI Pipeline Provenance Ledger Connectors Memory Stack ← Back to Overview