The third batch of submitted APIs is profiled and published — eight providers, 286 new artifacts. Six of the eight declare a home outside the English-speaking tech centres: Lithuania, Norway, Singapore, Saudi Arabia, Germany. The other two do not say where they are from at all.
| Provider | What it does | Where |
|---|---|---|
| Tvarka ATK API | Lithuanian eID authentication and qualified electronic signing, with timestamping and long-term validation | Lithuania |
| Atmospore | AI-driven global pollen and allergy forecasting, combining GFS weather data with professional aerobiology observations | Oslo |
| ShipFinder | Turns the global AIS feed into queryable vessel intelligence | Singapore (ELANE GLOBAL PTE. LTD.) |
| Sarj AI | Voice and chat AI agents built for Arabic, English and Urdu | Saudi Arabia / MENA |
| i6eal | 16 AI-related open datasets and monitors for Germany and the EU, free and keyless | Germany (Syka Ventures UG) |
| Autoderm | White-label API for AI-assisted analysis of dermatological images, which Autoderm operates as a regulated medical device | Europe |
| StayingAPI | Unified accommodation data — availability, cross-OTA price comparison and normalized reviews across Airbnb, Booking.com, Vrbo and Google Hotels | — |
| NutrientsDB | ~2.9M food entries across 86 normalized nutrient fields, deduplicated across 180+ countries, sold as a one-time dataset purchase | — |
National infrastructure is an API now
Tvarka is the one to sit with. It is a single REST API for Lithuanian national eID authentication and qualified electronic signatures — the legal instrument, not a signature image — plus timestamping and long-term validation so a signature still verifies a decade from now.
That is state-level identity infrastructure, exposed as an ordinary HTTP API by a company most of this audience has never heard of. The same story is playing out across the EU under eIDAS, and almost none of it is written about in English.
Atmospore is a quieter version of the same pattern: pollen forecasting sounds trivial until you notice it needs a global weather model fused with ground-truth observations from professional aerobiologists, and that it is the kind of environmental data layer that used to live only inside national meteorological institutes.
The regulated API
Autoderm is worth calling out for a different reason. It is a REST API you can integrate in an afternoon, and it is also a regulated medical device — a compliance surface most API providers never touch and most API tooling has no vocabulary for.
As the catalog accumulates providers under regimes like eIDAS, MDR and PSD2, the interesting question stops being “does it have an OpenAPI” and becomes “what is it legally allowed to do, and what did it have to prove to be allowed.” That is a gap in how the whole industry describes APIs.
On who wrote what
Every artifact carries provenance: searched when we found it published by the provider, probed when we confirmed it against a live endpoint, generated or derived when we authored it from what the provider does publish. This batch came out to 69 searched, 24 probed, 41 generated and 40 derived.
One submission in this batch produced no artifacts at all, and that is a feature. Its profile records an honest zero: no developer program, nothing machine-readable but an llms.txt, and every documented path answering 404 — including an apis.json that does not exist. We publish the evidence rather than model an API that isn’t there.
Every provider owns their record and can correct it — write to us and we will.
If your API is not in the catalog, add it.