POS
Speed. Keyboard-first. Zero latency.
Cloud restaurant OS, built for India
The restaurant OS that never drops the bill.
Offline-first billing, a kitchen that never loses a ticket, and AI that's native to the product — not bolted onto a dashboard.
Table view — live
T1, Fresh
T1Fresh
T2, Hot
T2Hot
T3, Warming
T3Warming
T4, Critical
T4Critical
T5, Warming
T5Warming
T6, Fresh
T6Fresh
Where restaurant software actually fails
Not on features. On the boring things.
A bill that rounds wrong.
Subtotal₹1,180.00
CGST 2.5%₹29.50
SGST 2.5%₹29.50
Payable₹1,239.00
Each tax component rounded half-up, independently. Never a float.
Money is bigint paise, never a float. CGST and SGST are each rounded half-up, independently — never 5% halved.
A KOT that prints twice.
KOT #0482Main Kitchen
- 2×Butter Chicken
- 1×Garlic Naan
- 3×Jeera Rice
reprinted × 1
KOT ≠ Bill: separate numbering, separate lifecycle. A reprint increments a counter — it never creates a second ticket.
A table stuck “occupied”.
T7 · LiveSplit → T7a / T7b · 0:04 ago
The floor is live state, not a stale flag — merge, split, and move read the same source the kitchen and the bill do.
And then the WiFi drops at 8:40 on a Saturday.
Offline-first, not offline-tolerant
Kill the network mid-service. Bill four tables. Reconnect.
Zero duplicate bills. Zero lost items. Zero duplicate KOTs. No unexplained invoice gap — the acceptance test we actually run.
Terminal
Local outbox
Reconnect
Server
Offline — writing to the outbox
One system, every surface
Six screens. One source of truth.
Floor
Table state at a glance — merge, split, move.
KDS
Readable at 2 metres, ticket aging, one gesture: bump.
Captain
Take order at table, fire KOT, call for bill.
Booth
Browse, get guided, order, pay, feedback.
Console
Menu, inventory, reports, AI insights, multi-outlet.
AI, native — not bolted on
Not a chatbot bolted onto a dashboard.
AI reads the same order, menu, and review data every other surface does — it never touches the ledger, and it never invents a number the schema can't back up.
Booth Host
A taste concierge at the table — guides the order, doesn't just take it.
Menu engineering
Reads real sales and margin data to say what to push, park, or cut.
Review → action
A bad review doesn't sit in an inbox — it becomes a task with an owner.
NL analytics
Ask the Console a question in plain English, get a real number back.
Built for the Indian F&B market
Not localised later. Native from day one.
Multi-outlet, multi-brand
One login, every outlet and every brand in a shared cloud kitchen — day one, not a v2 migration.
GST-correct, always
CGST and SGST computed independently and rounded half-up, on every invoice, at every outlet.
Zomato & Swiggy
Aggregator orders land in the same KDS queue as a walk-in table — no second screen to watch.
UPI, at the table
Pay at the table or at the counter — the invoice number never changes because of how it was paid.
Proof of craft
No logos. No pilot yet. Just what's actually true.
line and branch coverage on packages/domain — every money, tax, and KOT rule.
floats in the money path. Every paisa is a bigint, server-authoritative.
adversarial RLS test cases run as real Postgres roles, not mocked policies.
duplicate bills, lost items, or duplicate KOTs — the offline acceptance test.