Priya Life Science has launched Airmid Éireann, Ireland's open pharma data project. It publishes 19 free datasets with 502,489 records on clinical trials, medicines, shortages, approvals, safety alerts, trade, spending and research in Ireland and Europe. Every file names its source and licence, the data refreshes weekly, and the code is open source.
The project page is at priyalifescience.com/airmed, the files are at priyalifescience.com/opendata/airmid/v1, and the code is on GitHub. The first release went live on 13 October 2026 with 10 datasets and 32,976 records. Release 1.1, with data collected on 15 October 2026, added Europe-wide trials, EU trade and industry statistics, research projects and papers, and safety data.
- 19 open datasets, 502,489 records from 12 official publishers, including the HPRA, EMA, CSO, ClinicalTrials.gov, Eurostat, OECD, FDA, Health Canada and MHRA.
- Free to use: CSV and JSON files, a catalog.json index that works as a simple API, and a data dictionary for every field.
- Search by meaning across 184,076 records. The search model runs in your browser, so your search never leaves your device.
- Medicine cross-reference: 1,658 active substances that appear in two or more datasets.
- No personal data: no personal names or contact details, even when a source publishes them.
- Open and citable: code under EUPL-1.2 on GitHub, archived on Zenodo with the DOI 10.5281/zenodo.23336432.
1. What is Airmid Éireann?
Ireland makes medicines for the world, but the public facts about that industry are spread across dozens of registers, spreadsheets and reports, each with its own format and its own terms of reuse. Airmid Éireann collects those facts from the official publishers, checks them, and releases them as open datasets that anyone can download and reuse. It is the open-data project of Priya Life Science, and it uses the same public sources that power our registers and trackers.
Sreepriya, founder of Priya Life Science, sets out the aim on the project page: "Our mission is to gather Ireland's public pharma data, check it, say exactly where every number comes from, and make it free for anyone to use: students, researchers, journalists, patients and companies."
The name comes from early Irish mythology. Airmid was a healer of the Tuatha Dé Danann who, in the story, gathered the healing herbs that grew from her brother Miach's grave and sorted them on her cloak. The data catalogue is called Tobar Airmid, the well of Airmid, with the tagline "Ireland's well of open medicine data".
2. What data is in it?
Release 1.1 holds 19 datasets. Each comes as a CSV file and a JSON file, with a datapackage.json that records the source, licence, field descriptions and the date it was collected. The record counts below are from the live catalogue for data collected on 15 October 2026; they change slightly each week.
| Dataset | Records | Source | Licence |
|---|---|---|---|
| Clinical trials with an Irish site (all statuses) | 2,699 | ClinicalTrials.gov (U.S. National Library of Medicine) | Public data, with credit |
| Clinical trials with a European site (started 2015 or later) | 113,339 | ClinicalTrials.gov | Public data, with credit |
| Medicines authorised in Ireland | 10,294 | HPRA | CC BY 4.0 |
| EU centrally authorised medicines | 2,351 | European Medicines Agency | Reuse with credit |
| EU orphan designations | 3,310 | European Medicines Agency | Reuse with credit |
| EU medicine shortages | 85 | European Medicines Agency | Reuse with credit |
| US drug shortages | 1,585 | U.S. FDA (openFDA) | Public domain |
| US original approvals since 2021 | 4,772 | U.S. FDA (Drugs@FDA) | Public domain |
| Health Canada new drug approvals since 2016 | 1,060 | Health Canada | Open Government Licence (Canada) |
| UK medicine and device safety alerts | 2,323 | MHRA, via GOV.UK | Open Government Licence v3.0 |
| US medical device recalls since 2018 | 23,868 | U.S. FDA (openFDA) | Public domain (CC0) |
| US adverse event report counts (totals only) | 1,212 | U.S. FDA (FAERS, via openFDA) | Public domain (CC0) |
| Irish pharma exports by country | 6,811 | Central Statistics Office | CC BY 4.0 |
| EU pharma trade by country | 207,110 | Eurostat | Reuse with credit |
| EU pharma manufacturing by country | 32,436 | Eurostat | Reuse with credit |
| Pharmaceutical spending by country | 70,802 | OECD | Reuse with credit |
| EU-funded health research projects | 14,111 | CORDIS (European Commission) | CC BY 4.0 |
| Irish pharma research papers (no author names) | 4,300 | OpenAlex | CC0 |
| NIH awards to Irish organisations | 21 | U.S. National Institutes of Health | Public domain |
| Total | 502,489 | 12 publishers | Each dataset keeps its source licence |
Two of the datasets need a note. The FAERS counts are numbers of adverse event reports sent to the FDA; as the FDA itself says, a report does not prove that a medicine caused the event. The Eurostat data covers the EU member states only, because Eurostat's terms do not allow commercial reuse of its data on many countries outside the EU, EFTA and the candidate countries.
3. From Ireland to Europe
Airmid Éireann starts with Ireland: every trial with an Irish site, every medicine on the HPRA list, Irish pharma exports by destination, and research from Irish institutions. Release 1.1 adds the European picture. The European trials dataset covers studies started since 2015 with a site in any of the 27 EU member states, the UK, Norway, Iceland, Switzerland or Liechtenstein, and lists the European countries taking part in each one. The Eurostat datasets compare every EU member state's trade in medicinal and pharmaceutical products and the size of its pharma manufacturing sector, and the OECD dataset compares spending on medicines across 53 countries.
That makes it possible to answer questions such as how many trials Ireland hosts compared with countries of a similar size, or how Irish pharma exports compare with other member states. The same data can be explored on our registers, such as clinical trials in Ireland, medicines in Ireland, EU medicine shortages, Irish pharma exports and research funding to Ireland.
4. How to use it
There are four ways in, and none needs a sign-up or a key:
- Download the files. Every dataset is a CSV for spreadsheets and a JSON file for code, listed at /opendata/airmid/v1 with SHA-256 checksums.
- Read the catalogue from code. catalog.json lists every dataset, its licence, record count, collection date and file links. Any app can read it, from any website.
- Search by meaning. The search page finds records by what they are about, not only by exact words, across 184,076 trials, medicines, shortages, approvals, alerts, recalls, research projects and papers. The small search model downloads once (about 23 MB) and then runs in your browser, so your search never leaves your device.
- Follow a medicine across datasets. The medicine cross-reference lists 1,658 active substances that appear in two or more datasets. Methotrexate, for example, appears in seven: EU approvals, orphan designations, EU shortages, US approvals, US report counts, Health Canada approvals and the HPRA list.
For example, to list every dataset from the command line:
curl https://priyalifescience.com/opendata/airmid/v1/catalog.json5. How the data is checked
Every source has to be on an approved list before the collector will run it, with its licence page recorded. Each collector keeps only the fields named on that list. Where a publisher states how many records it holds, the count is checked, and a run that comes back short is refused rather than published half-finished. A new release is also refused if a dataset shrinks sharply without explanation. Every value passes a privacy filter that blanks anything that looks like an email address or phone number, so a source that starts publishing contact details is caught rather than copied.
The whole collection runs again every Monday on our own server. Each release keeps a changelog, and errors we correct are listed on our corrections log. How each register is built is explained on our methodology page.
6. What is never shared
Some data stays out of Airmid Éireann on purpose:
- Personal names and contact details. Investigators, officials and staff are never included. Some investigator-led trials list a person as lead sponsor; Airmid blanks those and keeps sponsors that are hospitals, universities and companies.
- Data licensed only for our own site. NICE appraisal facts, German shortage notices from BfArM and Japanese trial records from jRCT were licensed to Priya Life Science for our pages. They are not ours to pass on.
- Data with unclear terms. GMP certificates from EudraGMDP stay out, and so does the EU Clinical Trials Information System (CTIS), because the EMA's reuse terms do not cover data supplied by sponsors and member states. We are asking the EMA for written permission.
- Leaflet and guidance text. Patient leaflets, product information and guidance documents belong to their publishers, so we link to them instead.
7. Open source, and honest about AI
The code that collects and checks the data is open source under the European Union Public Licence (EUPL-1.2) on GitHub, with tests and secret scanning on every change. Each dataset keeps the licence of its source, and our own compiled data is released under CC BY 4.0. Every release is archived on Zenodo, so it can be cited with a permanent DOI: 10.5281/zenodo.23336432 for all versions.
Much of the code was written with an AI assistant, and the project says so in its README and in each commit that had AI help. The search uses a small open model, all-MiniLM-L6-v2, that runs on our own server to build the index and in your browser to read your query, so no data or searches go to an AI company. No value in the data is written by AI: models only help rank and link records, and the original record is always shown.
8. Who it is for
- Students and researchers who need clean, citable data for a thesis or paper, with the source and licence on every file.
- Journalists checking a claim about shortages, trials or exports against the official figures.
- Patient groups following trials, shortages or safety alerts for a condition.
- Companies and investors comparing Ireland with other European countries on trials, trade and manufacturing.
- Developers who want a stable, versioned feed of public pharma data without scraping a dozen websites.
9. What comes next
The roadmap on the project page lists the next datasets: patent and exclusivity dates, pharma manufacturing sites in Ireland, Irish legal entities of pharma groups, and medicine spending in the UK and US. EU trials from CTIS will follow if the EMA gives permission. New sources join only after their licence has been checked, and suggestions are welcome through the GitHub issue forms or by email to info@priyalifescience.com.
10. Frequently Asked Questions (FAQ)
Q: Is Airmid Éireann free to use?
A: Yes. All 19 datasets are free to download and use, with no sign-up. Each dataset keeps its source's licence; most allow commercial reuse with credit to the source and to Airmid Éireann.
Q: Where does the data come from?
A: From 12 official publishers, including the HPRA, the European Medicines Agency, the Central Statistics Office, ClinicalTrials.gov, Eurostat, the OECD, the U.S. FDA, Health Canada, the MHRA, the European Commission's CORDIS service and OpenAlex. Every file names its source, licence and collection date.
Q: How often is the data updated?
A: Every week, on Mondays. The catalog.json file always shows the current record counts and collection dates.
Q: How do I cite Airmid Éireann?
A: Use the DOI 10.5281/zenodo.23336432, which covers all versions, and credit the original source of the dataset you use. BibTeX and RIS files are in the release folder.
Q: Does the search send my query to an AI company?
A: No. The search model downloads to your browser and runs there, so your query stays on your device.
Q: Is this medical advice?
A: No. Airmid Éireann publishes public records for research and information. Always check the original source, and speak to a healthcare professional about any medical decision.
Conclusion and Outlook
Airmid Éireann turns Ireland's scattered public pharma records into one open, checked and citable collection: 19 datasets and 502,489 records today, refreshed every week, with the code open for anyone to inspect. The next steps are more Irish data, starting with patents, manufacturing sites and company identifiers, and wider European coverage as permissions allow. Explore it at priyalifescience.com/airmed, and tell us if you find an error or know a source we should add.
Sources: Airmid Éireann release catalogue (catalog.json, data collected 15 October 2026); Airmid Éireann project page and repository (github.com/priyalifescience/airmed); Zenodo record 10.5281/zenodo.23336432; the licence pages of each publisher listed in the release. This article is for information only and is not medical advice.
Cover image: AI-generated illustration of an Irish holy well, not a photograph of a specific place. Inline illustrations: Airmid Éireann project artwork.
