1. Current subprocessors
| Provider | What it does | Location |
|---|---|---|
| Amazon Web Services | Hosting, delivery, storage, text translation (Amazon Translate, Vocaic’s main translator), and the AI models that write Ava’s answers and, when other translation is unavailable, translate a message | United States |
| Microsoft | Speech synthesis for spoken messages | United States |
| Translated srl (MyMemory) | Text translation for the languages Amazon Translate does not cover, and as a backup | Italy |
| Wikimedia Foundation (Wikipedia) | Receives a short search query when Ava looks up a public fact, and returns the public articles she cites | United States |
| Your browser vendor | Live speech-to-text in your browser (for example Google in Chrome, Microsoft in Edge, Apple in Safari), under that vendor’s terms | Varies by vendor |
Transcription in the Accent Studio runs an open-source speech model on Vocaic’s own server, so it is not a subprocessor. A Slack, WhatsApp, Telegram or Discord app, or an Alexa skill, that someone connects to Vocaic is their own service, used under their own agreement with that platform; to check Alexa requests, Vocaic fetches Amazon’s public certificate (no personal data is sent).
2. Data processing agreements
Business customers who need a data processing agreement, or who want to be notified of changes to this list, can contact agarw574@purdue.edu. This page will be updated before a new subprocessor begins handling personal information. See also the Privacy Notice and the Security page.