FieldOps AI
The system that proves work happened.
GPS-verified clock-ins, photos stamped by the server, and a live supervisor view, replacing WhatsApp photos and Excel timesheets for construction, electrical and infrastructure crews.
What we delivered: Product, Flutter mobile app, supervisor web app, backend
Good morning, Diego
Riverside Substation
Live timeline
Realtime- Tap the phone to clock in. Try going offline first: events queue on the device and sync when the signal comes back.
Events are appended, never edited. Each one carries server time, GPS and the worker who did it.
Interactive preview rebuilt from the real FieldOps AI screen, with sample data. Not connected to production.
- verified location on every clock-in and photo
- GPS
- verified location on every clock-in and photo
- roles: worker, foreman, supervisor, admin
- 4
- roles: worker, foreman, supervisor, admin
- edits or deletes in the event store
- 0
- edits or deletes in the event store
- first, with automatic sync
- Offline
- first, with automatic sync
Figures from the FieldOpsAi README.
The problem
Field companies lose money on disputes they can't settle: unpaid invoices, payroll errors, “nobody told me.” The cause is always the same. There is no proof of what happened on site.
What we built
GPS-verified clock in and out
Server timestamp and device GPS are locked at the moment of action, in one tap, in under 30 seconds.
Photos stamped by the server
Worker, GPS, time, job reference and a verification hash are burned into the image pixels, so a photo carries its own evidence.
Offline-first mobile
The Flutter app works with no signal and syncs through a local queue when the crew is back online.
Live supervisor dashboard
Supabase Realtime powers the map, photo feed and job timeline with no polling.
An append-only event store
Events land in partitioned Postgres and are never edited or deleted, with row-level security per company.
Overtime and reports
Overtime requests are routed and approved with a full audit trail. Timesheets and job summaries export to PDF for billing.
Architecture
How it fits together
Crew
- Flutter app (worker + foreman)
- Offline queue (Drift)
Backend
- Supabase Edge Functions
- Photo stamping
- Append-only event store
Data
- Postgres 17, partitioned
- Row-level security per tenant
Office
- Next.js supervisor dashboard
- Realtime map + timeline
- PDF reports
Outcomes
- 80%less field-to-office communication lag (Team-reported)
- 12 hrs/wkof manual dispatch tracking removed (Team-reported)
Want something like this?
Next project
Keniya
E-commerce store
A condition-aware online store with a review gate.