How Sprig uses it
Sprig uses this data to run the service: authenticate your account, capture and store recordings, produce transcripts, separate and name voices, create recaps and embeddings, search your meetings, and show the results back to your workspace.
Where data is processed
Sprig stores meeting, transcript, and account data in its Postgres database, and meeting audio in private object storage. Sprig uses these providers to deliver the service:
- Clerk — authentication and account identity.
- Supabase — the Postgres database holding meetings, transcripts and account records.
- Cloudflare — private object storage (R2) for meeting audio, attachments and exports.
- Deepgram — speech to text.
- ElevenLabs and Cartesia — speech to text, used only as fallbacks if Deepgram is unavailable mid-meeting.
- OpenAI — summaries and embeddings.
- Railway — API, realtime, and worker hosting.
- Vercel — web hosting.
- PostHog — product usage and web performance analytics.
Product analytics
Sprig records page visits, browser performance measurements, and a small set of product milestones such as starting a recording or completing a recap. Analytics uses pseudonymous identifiers. Sprig does not send meeting audio, transcript text, meeting titles, raw account or workspace identifiers, or meeting identifiers to PostHog, and does not use PostHog session replay or automatic click capture.
Retention and deletion
Audio and transcripts are retained until you delete them. There is no automatic retention window. Deleting a meeting requests removal of its recording, transcript, and recap; end-to-end storage deletion verification and an audit workflow are still being completed.
See also the Terms.