Kurdify is a Sorani-first business directory and commerce platform for the Kurdistan Region of Iraq, with a second lane into the global Kurdish diaspora. 13,271 businesses and 261,294 reviews across 12 cities are live in the product today. It is pre-revenue, and this page explains exactly what exists, what it runs on, and what the money changes.
Not a prototype and not a deck mockup. The rails below are running in production and can be opened in a browser right now.
13,271 businesses, 261,294 reviews, 12 cities, collected and cleaned by an in-house pipeline, with a server-rendered page per business carrying structured data for search engines and answer engines. This is the asset that takes the longest to build and it already exists.
A 25-tab owner dashboard: profile, hours, catalogue, promotions, coupons, jobs, events, reviews, billing and performance. Plus a claim-and-verify flow so an owner takes control of a listing that was already there.
Directory, magazine, jobs, marketplace, real estate, autos, a buyer-requests board and a creator marketplace, with a points and Credit loyalty economy tying usage to discounts at verified merchants.
A per-business AI agent on the WhatsApp Cloud API that answers customers, books appointments into a calendar and meters its own message cost. $200 a year plus usage. WhatsApp is where commerce in this market actually happens.
A Kurdish-language assistant with voice in and voice out, plus an advisor that answers a business owner's questions from its own data, industry data and market data as three separated sources rather than one blended guess.
Placement is decided by an online learning model with exploration, over signals like category and city match, rating, review volume, recency and click history. It is a statistical ranker, deliberately not dressed up as anything larger.
Four public surfaces, one platform, one dataset. A visitor, a business owner, a WhatsApp customer and a search engine all reach the same source of truth.
High-level request path. Everything to the right of the edge is Kurdify's own code.
Built for a market where a fixed monthly cloud bill is a real constraint, so the engineering choices are deliberately unfashionable and deliberately cheap to run.
Why this matters commercially
The whole platform runs at near-zero fixed infrastructure cost. That is not an aesthetic preference, it is the reason a business can be charged $450 a year in a market where a Western SaaS price is impossible, and still leave a margin. The engineering constraint and the pricing strategy are the same decision.
The commercial mechanism, end to end. Every business in the directory starts at step one already, which is the point of building the dataset first.
The owner funnel. Gold marks a step where money or a commitment changes hands.
The loyalty economy sits underneath all four: consumers earn points for real activity, points convert to Credit, and Credit is redeemed by QR at a verified merchant. The merchant gets measurable footfall, which is the proof that renews the subscription.
This is the honest part, and it is the clearest single argument for the raise.
The application currently runs on one machine, published through a secure tunnel. It is supervised, it restarts itself, it is backed up offsite nightly with verified restores, and its public reachability is checked every five minutes. It is a genuinely careful setup, and it is still one machine.
There is no second origin, no failover, and no alerting that reaches a phone. That is an acceptable cost while the product is pre-launch with no customers depending on it. It stops being acceptable on the day the first hundred businesses are paying.
Every dollar not spent on infrastructure went into the dataset and the product. 13,271 businesses and 261,294 reviews were collected, cleaned and made searchable without a hosting bill, by one engineer, before raising anything.
The architecture was chosen so that moving to real infrastructure is a deployment change and not a rewrite. The application is already containerised, the database is a single portable file, and the media layer can be pointed at object storage with one setting.
What an investor should take from this
The risk here is operational, not architectural. There is no technical debt to unwind and no rebuild in the plan: the hard part, a clean data model and a working product, is done. What is missing is the boring, purchasable layer, which is hosting, redundancy, monitoring and a second pair of hands. That is precisely what the infrastructure and hiring lines of the raise pay for.
$500,000 over 18 months on a YC post-money SAFE at a $6,500,000 cap, roughly 7.7% dilution. The allocation below is the one in the business plan, not a summary written for this page.
| Category | Share | Amount | What it actually does |
|---|---|---|---|
| Go to market | 35% | $175k | Ground sales in Erbil and Sulaymaniyah, paid acquisition, influencers and events. The dataset exists; this converts it into claimed and paying businesses. |
| Product | 30% | $150k | Native mobile apps, the Credit economy, and the diaspora lane. Mobile is the single biggest product gap today. |
| First hires | 20% | $100k | A Kurdistan team: sales lead and two reps, a Sorani community and content manager, support and operations, and a contract mobile engineer. |
| Legal and entity | 8% | $40k | Holding company structure, securities counsel per jurisdiction, local operating company, and the compliance work the loyalty and prize mechanics require. |
| Infrastructure and reserve | 7% | $35k | Off the single machine and onto managed hosting with redundancy, monitoring and alerting, plus a reserve. |
On the projections
Every figure above is a target, not a commitment, and the platform is pre-revenue today with one business on a paid plan. The base case exits month 18 at roughly $380k to $450k annualised. The long-run reference point is the Kurdistan Region alone: 200,000+ businesses, and 2% of them at an average of $40 a month is about $2M in annual recurring revenue, before any transaction income and before the diaspora lane.
Next
This page covers the technology. The market case, the model and the full data room sit alongside it.