Sub-processors
These are the third parties that process customer data on our behalf. We notify customers at least 30 days before we add one or change what an existing one receives, and a customer may object within that period. Our data processing agreement gives each one's legal entity and the basis for any transfer outside the EEA.
Processing article content
These receive the text of articles or drafts, or extracts from it.
| Service | Purpose | What it receives | Location |
|---|---|---|---|
| Microsoft (Azure OpenAI) | Language models for analysis, Find sources and archive search | The full article or draft and the outlet profile | EU Data Zone, Sweden |
| Google Cloud | Runs our servers; stores uploaded files, reports and database backups; translates Danish and Norwegian articles before batch scoring | Everything the service processes | Netherlands. Translation in the EU |
| Supabase | Database and sign-in | Accounts, analysis results including the passages quoted from your text, article extracts | Germany. The company is based in Singapore |
| Lettermint | Email: drafts sent by email, the reply and account emails | A draft sent by email, the reply, and recipient addresses | NetherlandsEvery message is stored for 28 days |
| Sentry | Error monitoring | Error reports, configured not to include article text | Germany. The company is based in the United States |
Also worth knowing
Find sources sends search queries written from your article to Serper.dev and OpenAlex in the United States. People's names are taken out of a query before it is sent, though in rare cases one can get through. Neither service has a data processing agreement with us, and neither is a sub-processor. Find sources is off until a journalist turns it on for themselves, and the account owner can switch it off for the whole newsroom.
Microsoft may store text its systems flag, for up to 30 days, and have it reviewed by staff in the EEA.
Questions about data handling