Every week tells me they're vibe-coding an alternative to Open Opportunities
- Ian Makgill
- Software
- 16 Oct, 2026
- 02 Mins read
Every week someone tells me they're going to vibe code their own version of Open Opportunities.
Genuinely, I wish them the best, but if you don't want to use our API, here's your task list...
- 1,050+ sources in 180+ countries, collected twice a day
- A scraper for each source, that is monitored on every run
- The UK alone has hundreds of portals, Find a Tender, Contracts Finder, Public Contracts Scotland, Sell2Wales, eTendersNI are the table stakes.
- Building a search that relies on activity not keywords
- Validating every record before it goes live
- Searching in any language: Tiefbau, génie civil, ingeniería civil
- Test suites for search - 94% of relevant tenders found across 900 searches on 9 portals
- Checking every field of every source every day, and publishing the results, errors included
- Adding a new source when a customer asks
- Ability to read “10.274.880,00 lei” as a number
- Converting every value to GBP, USD and EUR
- Detecting the language of every notice
- Spotting where the tender came from, country, county, city the lot.
- Adding scored CPV codes to every notice with 95% accuracy
- Spotting and sorting duplicates.
- Whilst never merging “carriageway repairs on A27” with the same request for the “A26”
- Finding and storing tender documents, wherever they can be gathered
- Search by pasting in a URL
- Translating any notice
- A button that turns walls of badly formatted text into readable prose
- Every award published since January 2021
- Knowing a tender closing in 2029 is usually a dynamic purchasing system, not a typo
- Saved searches and a digest in your inbox each morning
- CSV exports, bookmarks and a history of every search you’ve run (productivity tools)
- The original source link on every notice
That's before I mention making sure our API and the queries it can handle, our MCP server (coming soon), research reports, daily data quality reports, or the side-by-side test where we run your searches against ours and show you what each of us missed.
Oh, and besides maintaining your scrapers on your own forever, your vibe-coded version will still need you to pay for and stitch together:
- Headless browsers and proxies
- Hosting
- Database and search
- LLMs for classification and extraction
- OCR for scanned documents
- Translation
- Exchange rates
- Storage
- Email digests
- Auth
- Payments
- Monitoring
- Analytics
- Backups
- Archiving
