Get Started

Launch a micro-SaaS on UK open data and charge monthly

In ten minutes you'll have every food business in your council rated 0, 1 or 2 for hygiene, pulled live from a free government API: a list that sells for £19 a month. Then build the alert engine, Stripe subscriptions and a £5 server. The code is the easy half; this mission also covers the weeks of outreach that land the first customer.

315 minIntermediateTools: Free to build and test. With customers: about £5 to £6 a month for a server, a domain, and Stripe fees of roughly 70p on each £19 paymentStep 1 free
A working alert enginea tested Python script that pulls one council's food hygiene ratings from the FSA's free API, stores them in SQLite, detects new and changed ratings and renders the alert email
The licence rules for nine UK datasetswhat the Open Government Licence lets you sell, the exact attribution lines, and the traps (Royal Mail address data, FSA trademarks, personal data)
Monthly billing that runs itselfStripe Checkout, a signed webhook that keeps your subscriber list in step, and the Customer Portal for card changes and cancellations, with offline tests
  • A kit: the scripts, a dataset table, a SaaS maths calculator, outreach messages, onboarding questions, terms and privacy outlines, a scoreboard and a 30-day plan
Before you start
  • Python 3.10 or newer and uv. Install uv with winget install --id astral-sh.uv -e on Windows, or curl -LsSf https://astral.sh/uv/install.sh | sh on a Mac or Linux (Make a command-line tool walks through it). You should be comfortable running a script and reading an error
  • Helpful first: Build your own API in Python with FastAPI and Run n8n 24/7 on a £5 server (for the server set-up)
  • One niche you understand, or can learn quickly by talking to people in it
  • About 8 to 10 hours a week for the first two months: roughly half building, half talking to potential customers
01/08Why boring government data is a real product0 done
STEP 01 of 08~25 min

Why boring government data is a real product

See why people pay for free data, who already sells it, and what it takes before you write any code.

  1. The data is free. Watching it isn'tA food safety consultant wants to know the day a takeaway in their town drops to a hygiene rating of 1. An estate agent wants to know what sold on their streets last month. A letting agent wants to know when the Environment Agency issues a flood warning near their properties. All of that is published, for free, under the Open Government Licence ↗, which allows commercial use. What nobody gives them is a short, clean email that says "these three things changed, here's the link". That email is your product.

  2. People already pay for this shape of thingPlanning Signal charges £49 + VAT a month for planning application alerts (planningsignal.co.uk ↗). Same idea, different dataset: public records, filtered for one kind of buyer, delivered on time.

  3. And free tools exist tooHygieneScout sends free rating-change alerts to the public (hygienescout.co.uk ↗), and drivers can already get a free MOT reminder from gov.uk ↗. So "alerts about X" isn't enough on its own. What you sell is a business outcome for one kind of buyer: the right area, the right threshold, the right day, in a format they can act on in a minute.

  4. What it takes:

    • Micro-SaaS is a sales job with some code attached. The engine in this mission is an afternoon's work. Finding people who'll pay takes weeks, so flip the usual order: sell first (step 4), then build the app.
    • Time to the first paying customer: plan on four to eight weeks of steady outreach (around 60 personal messages a week) once you have a working sample. Asking is how you find out which niches convert.
    • Price for volume. At £19 a month, 50 customers is about £900 a month before tax. Step 8 shows the maths, including how cancellations set a ceiling on growth.
    • Check the data every week. Ratings appear online only after the business has had time to appeal. Land Registry sales show up weeks or months after completion. Councils fix and re-issue records. A weekly spot-check means you catch quirks before your customers do.
    • Read the licence. Most of the data is OGL, but address fields, logos and personal data often come with extra conditions. Step 2 shows where they apply.
    • Budget for running it. Even a tiny SaaS needs someone answering emails, renewing domains and noticing when a government API changes: 2 to 4 hours a week once it's running.
    TipWho this suits: you enjoy talking to customers as much as building, you can give it about 8 hours a week for a couple of months, and you're happy to sell by hand before any code does it for you. Keep your day job until paying customers prove the idea. Everything in steps 1 to 5 is free, so there's nothing to spend upfront.
  5. The kitDownload everything at once: the kit (zip). Or open the pieces: kit.md (niche picker, licence checklist, offer, messages, onboarding, terms and privacy outlines, checklists, routine, 30-day plan, prompts), fhrs_alerts.py, app.py, send_alerts.py, test_app.py, ppd_street_watch.sql, datasets.csv, saas-maths.csv and scoreboard.csv.

✓
CheckpointYou've written down two or three kinds of business you could sell to, and blocked out the hours a week you can give it.
If it breaksIf the data is free, why wouldn't they just look it up? + 1 more
If the data is free, why wouldn't they just look it up?
Same reason people pay for a cleaner when they own a mop. The FSA site lets you search; it doesn't tell you what changed this week in your patch. Your customer's time is worth more than £19 a month, and one useful alert pays for the year.
Is it legal to charge for government data?
Under the Open Government Licence, yes: commercial use is allowed, with attribution, and without suggesting the publisher endorses you. Some datasets add conditions (step 2). You're charging for the filtering, cleaning and delivery, not the data itself.
Still stuck? Ask the community
Next mission · Make money with AIFind the public data nobody has cleaned, and sell it as a paid API
→