Opening the kitchen

R&D active · Helsinki design partners

Your kitchen runs on assumptions.Orbis runs on data.

The operating intelligence restaurants have never had. Orbis connects what you'll sell to what you prep, order and staff, and shows its reasoning before the week starts.

Apply for R&D access →

Scroll · the sketch becomes data

1The guest

Fri · 16:30

One Friday service · follow Mikko

It starts before anyone walks in.

It's raining on the street outside. Mikko and a friend decide on burgers before the concert down the road. Aino, who owns the place, is already ready for them. She just doesn't know their names yet.

Behind the scenes · Orbis

  1. 01Reads the rain, 38 bookings, the concert 1.2 km away and the last four Fridays
  2. 02Forecasts 262 covers tonight, 9% above a usual Friday
  3. 03Stretches the prep and puts one more server on from 18:00

2POS

Fri · 19:42

The POS takes the order. Orbis reads the signal.

Mikko orders two Royal Burgers, no onion, fries and a bottle of the house red. The card goes through. Nothing else happens on screen. Everything else happens underneath.

Behind the scenes · Orbis

  1. 01Records the sale as an event that can't be edited, even if the Wi-Fi drops
  2. 02Compares live covers with the forecast: 188 so far against 181 expected
  3. 03Tells the kitchen, stock, labour and margin in the same second

3Kitchen

Fri · 19:42

The ticket fires to the line, and the recipe explodes.

The ticket prints at the grill and two patties hit the flat-top. Nobody on the line notices the recipe graph quietly taking 40 g of truffle mayo off the batch Jussi made this morning.

Behind the scenes · Orbis

  1. 01Routes the ticket to the grill and fry stations
  2. 02Breaks each dish into what it really uses, including house-made preps
  3. 03Tracks the mayo batch: 0.6 L left, enough until about 21:30

4Walk-in

Fri · 20:15

Every plate moves the shelf.

Behind the walk-in door, 3 kg of chicken from Tuesday's delivery hits its internal use-by at 18:00 tomorrow. Nobody has to open the door to check.

Behind the scenes · Orbis

  1. 01Deducts stock by recipe as it sells, oldest lot first
  2. 02Logs the walk-in at 3.1 °C every 15 minutes
  3. 03Flags the chicken with five options and suggests it for tomorrow's special

Without Orbis · the chicken is found at close, in the bin.

5Misa

Fri · 22:40

Tomorrow's prep is planned before tonight ends.

Service is over. Aino wipes down the pass. Tomorrow's prep board is already on her phone: 42 meatballs, 18 litres of soup, 31 salmon portions. She changes two numbers, taps confirm and goes home.

Behind the scenes · Orbis

  1. 01Turns tomorrow's forecast into quantities per station
  2. 02Nets off what's already made and keeps everything at or above par
  3. 03Locks the stock so no other plan can take it

Without Orbis · forty portions of salmon get prepped for a Tuesday that never comes.

6Supplier

Sat · 07:30

The order writes itself. You approve it.

The Nordic Fresh van backs into the alley. The order it carries wrote itself last night and Aino approved it with one tap. The courier hands over the crates and the probe reads 2 °C.

Behind the scenes · Orbis

  1. 01Builds the cart from forecast, par and shelf stock, in the supplier's pack sizes
  2. 02Prices it from Aino's own invoice history
  3. 03Checks quantity, temperature and use-by at the door, then books the lot into stock

Without Orbis · the invoice lands and you have overbought again. Or you run out at 19:40.

7Team

Sat · 09:00

The rota fits the forecast, and the law.

Saturday's rota went out on Wednesday. Three on the floor until 22:00, three on the line, evening rates from 18:00, and Jussi's Sunday premium already counted for tomorrow.

Behind the scenes · Orbis

  1. 01Turns forecast covers into staff hours for every hour of the day
  2. 02Drafts a rota that is TES-legal by construction
  3. 03Shows labour as a share of revenue before the shift starts: 27.4%

Without Orbis · five are on the floor for a service that needed three, and TES maths is done by hand.

8Orbis

All weekend

One restaurant. One state. Every decision explained.

One weekend, eleven stations, thousands of events, one picture of the restaurant. Aino doesn't open seven apps. She opens one, sees the four things that need her, and decides.

Behind the scenes · Orbis

  1. 01Keeps one shared state of sales, stock, prep, labour and margin
  2. 02Surfaces only what needs a decision, with the reasoning attached
  3. 03Learns from every approval and every override

Try it

Change the day. Watch the kitchen replan.

Day

Weather

Nearby

Tables booked

48

Shadow mode · live in R&D

Before Orbis acts, it watches.

Orbis runs a shadow copy of your restaurant beside the real one, predicting every service and checking itself against what actually happens. You decide when to let it act.

0%

covers forecast accuracy by week 3 (20% → 54% → 84%)

Orbis runs silently. You see its results against actuals.

For those who want to know why

Features can be cloned. This can't.

Every feature we ship first was chosen because it captures the data a later feature needs. The loop isn't just a product decision. It's how the data accumulates, and that advantage compounds.

Now

Finnish TES, encoded

The MaRa collective agreement in code: Sunday premiums, evening rates, three-week averaging. No international competitor has built this.

Year 1–3

Per-kitchen learning

Every forecast correction, misa override and supplier choice is a training event. The model in your kitchen at month 18 can't be copied by anyone starting today.

Year 3+

Network effects

Enough kitchens sharing anonymised patterns means a new restaurant gets a working forecast on day one, plus supplier price intelligence from pooled invoices.

Roadmap

What's live. What's next.

Orbis is in active R&D with Helsinki design partners. This is the honest picture, with no horizons squashed into one feature list.

H1 · R&D now

MVP, testable Dec/Jan

  • POS integration, read-only with history backfill
  • Recipe graph, from dish to ingredients to cost
  • Demand forecast in shadow mode
  • Misa list with reasoning shown
  • Recipe-locked inventory deduction
  • Draft purchase cart per supplier
  • TES labour cost view

H2 · Year 1

After soft launch

  • One-tap supplier ordering
  • Prep sequencing by station and time
  • Waste vs theoretical variance
  • Demand-driven, TES-legal shift scheduling
  • Payroll export to Netvisor / Procountor
  • Menu engineering by margin and velocity
  • Oiva food-safety compliance export

H3 · Year 2–3

The data flywheel

  • Network cold-start: a forecast on day one
  • Supplier price intelligence from pooled invoices
  • Autonomous replenishment for stable ingredients
  • Compliance plug-ins for Sweden, Estonia, UK
  • Trend radar for rising dishes nearby
  • Revenue simulation: what if we open Mondays?

The CodFleet ecosystem

Orbis is the first layer. The others sharpen it.

CODOrbis

Operating intelligence

Forecast, prep, ordering, labour and TES-legal payroll in one loop: rotas, time and pay for your whole team.

The forecast learns from

Live event logsWeatherFootfallOffers & promotionsCODGo orders & bookings

CODGo

Consumer app · orders · bookings

The guest app for ordering food, booking tables and booking home services. Every order and booking also sharpens the forecast.

CODAlly

Workforce network

Couriers, cleaners and service providers, ready when a restaurant or a home needs them.

That was one Friday service

Now picture it in your kitchen.

Apply for R&D access →

Less guessing. Better margins.

See how Orbis would run the loop for your kitchen, in shadow mode first, with no risk to service.

R&D cohort · limited places

Apply for early access.

We're onboarding a deliberate R&D cohort: owner-run kitchens, franchise operators and multi-site chains. Whether you run one room or fifty, if you make the decisions about what to prep, order and staff, Orbis is built for you.

  • →Independent restaurants, franchises and chains
  • →An owner or operator involved day to day
  • →An existing POS (any major provider)
  • →Willing to share feedback during R&D
  • →Helsinki preferred, other cities considered

R&D access is selective and free during testing. We read every application personally and reply within 48 hours.