Food is the largest system humans run together, and very little of the value inside it reaches the people who cook. A person who earns reliably from food can feed their own household, and keeps feeding it. A food system only lasts if the people inside it can afford to stay in it.
Where we are today
Launching in Helsinki in September.
€138Kpre-seed raised to date
15,000+people on the waiting list
400+food creators applied to join
Helsinkifirst city, rollout from September
One part of this has been tested in the real world: physical food ordering, in Turku. Everything else goes live from September, beginning with food and recipes in Helsinki and digital products everywhere.
Kokit is in Batch 3 of the Northern Light Accelerator at Maria 01. Our founder was selected for the founding executive cohort of the Chicago Booth and OpenAI MasterClass programme, from more than 29,000 applicants, learning how AI native products are built from the people who build them best.
What we do not claim yet: creators earning on the platform, meals sold, waste avoided. Those figures begin at launch. The 2026 baseline is where our results start.
The scale we are building toward
What 2035 looks like if we get this right.
1M
food creators earning on Kokit
150K
of them making a full living from it
20M
people buying from them
150+
food cultures earning
And what that produces every year.
26Mhours of administrative work returned to creators every month6 hours a week each, 312 hours a year, 39 working days
15,000 tfood waste avoided every year153M cook-to-order meals, 100 g avoided per meal
38,000 tCO2e avoided every year2.5 t CO2e per tonne of food not wasted
3+income streams the average creator learns to runthe skill the agent teaches while it works
How these are calculated. Hours: our 6 hour weekly target against a measured baseline of 11 hours a week of admin for small business owners. Food waste: 55% of turnover in physical meals at an €18 average order, 100 g of overproduction avoided per confirmed order. Independent studies find roughly 82% of restaurant food waste happens at preparation and service, before a plate ever reaches anyone, which is the portion cooking to confirmed orders removes. Carbon: the widely used ratio of about 2.5 tonnes of CO2e for every tonne of food wasted. These are modelled projections at full scale. We will publish the measured figures from the 2026 baseline onward.
The path there
Drag the year.
Drag to move through the years
20262035
2026
Helsinki
400food creators
launchingmaking a full living
6,000people buying
10,400hours returned monthly
6 tfood waste avoided yearly
15 tCO2e avoided yearly
2026 to 2028 is our funded plan: Helsinki, then eighteen European cities. From 2029 the figures are the global scenario, and they depend on further investment.
How we ask to be judged
Three tests.
Test one
Impact
Did real lives measurably change, and can the number be checked by somebody outside Kokit?
Test two
Innovation
What could not be done before Kokit existed?
Test three
Inspiration
Can others copy it, build on it, and hold us to it?
Honest take
Where this comes from, and what is still thin.
Kokit was built by a trained chef who spent a decade in food, through restaurants that closed and ventures that did not work, building the thing he needed himself ten years ago. The person the system overlooks was the first user we designed for.
What is thin is delivered impact. We are at launch stage, so almost everything below is a commitment. We work on food security through income, waste and distance. Direct hunger relief sits outside what we do.
The 2035 figures are goals we are setting, and they reach beyond our funded twenty four month plan. Where a number is an estimate, we say so.
The commitments
Ten things we commit to build by 2035.
Each one carries a target, and the method we will be measured by. They are grouped under the test each has to pass.
Impact
Commitments 01 to 06
01
Return the majority of every sale to the person who made the food
Eighty percent of every sale goes to the creator the moment it sells. The twenty we keep is what keeps Kokit running. Nothing to pay to start, so we only earn when a creator does.
Target by 2035150,000 food creators earning a full living, with at least 80% of every sale kept by the creator
How we measure itShare of gross sale returned to creators, and the number earning above a local living wage, published yearly
Where we start: more than half of all creators earn under $15,000 a year, and only around four in a hundred pass $100,000. The share is where the gap sits.
02
Treat food creators as professionals
Dignity is structural. Creators set their own prices, choose what they sell and to whom, own their recipes and their content, see exactly what they are paid and when, and can appeal any decision that affects their account. Our terms are written to be read.
Target by 2035A published fair terms standard covering pricing freedom, content ownership, payout clarity and appeal rights, in force in every market
How we measure itPayout speed, appeal outcomes, and an annual creator confidence survey published unedited
03
Make it possible to start with almost nothing
A food business has always demanded a lease, equipment, staff and a year of nerve. On Kokit the first step is one product, sold from the kitchen someone already has. Everything else can come later, at their pace.
Target by 2035The clear majority of new food creators start earning with no upfront capital and no premises
How we measure itShare of new creators making a first sale without capital outlay, and time from joining to first payment
Context: there are roughly 23 million food service establishments worldwide, and far more people who could earn from food but have never had a route in.
04
Open the door to the people the system overlooks
Some of the best food in any city is made by people the formal economy struggles to place. Newcomers to a country. People working in a second or third language. Carers who cannot leave the house for a fixed shift. We build for them deliberately, through language coverage, plain onboarding, and guidance on local requirements.
Target by 2035Onboarding and support in 40 or more languages across 40 or more countries
How we measure itLanguages supported, and the share of active creators onboarding in a language other than the local majority one
Food is one of the few trades where skill crosses a border intact. A recipe needs no recognition of qualifications. It needs a door.
05
Cut food waste at the source
Most food waste is decided before anyone sits down to eat, by someone guessing how much to prepare. On Kokit the order usually exists before the cooking starts. Forty portions already sold waste close to nothing, and the model rewards accuracy. We will measure the waste avoided per order and publish it.
Target by 2035Waste avoided measured on every order and published annually, aligned with the goal of halving per-person food waste
How we measure itPortions sold before preparation as a share of all orders, and estimated waste avoided per order against a local baseline
Context: the EU generates over 59 million tonnes of food waste a year, about 132 kg per person, worth roughly €132 billion. Food services account for around 15 kg of that per person. Globally the figure passes a billion tonnes. UN SDG 12.3 calls for halving per-capita food waste by 2030.
06
Shorten the distance between the kitchen and the table
A meal bought straight from the person who made it carries a shorter chain behind it. Kokit is local by design, and digital products such as recipes and classes travel with no physical footprint. Shorter chains also mean more resilient local supply, which matters most where the long chain breaks first.
Target by 2035A published footprint-per-order figure with an annual reduction target, working toward the Paris-aligned benchmark for an average meal
How we measure itEstimated emissions per fulfilled order, split between physical and digital products, in an annual climate report from 2027
Benchmark: WWF has set out that the average meal needs to reach roughly 0.5 kg of CO2e to stay within a 1.5°C pathway. We will report against that figure.
Innovation
Commitments 07 and 08
07
Give creators their hours back
Each creator's AI agent takes on the administrative work that fills their evenings. Human time is a resource, and it is the one we are most determined to protect.
Target by 2035At least 6 hours a week returned to every active creator, over 300 hours a year each, measured and published
How we measure itTasks completed by the agent instead of the creator, translated to average time saved
08
Make income reliable, not just possible
Selling one product occasionally does not build security. The agent learns what sells and helps creators add new streams - moving from a recipe to a class, or adding a subscription - so they are not dependent on a single source of income.
Target by 2035The average creator running 3 or more distinct income streams on the platform
How we measure itAverage active product types per earning creator
