Last updated: October 8, 2026
SolScribe is a zero-retention cloud transcription service. You upload audio; we transcribe it in an isolated GPU environment; we delete the staged audio when the job completes. Your transcript stays in your account, encrypted at rest, until you delete it. We never use your audio or transcripts to train AI models.
This policy covers the hosted service at app.solscribe.ai and the marketing site at solscribe.ai. SolScribe's self-hosted edition runs on your own hardware, and The Self-Hosted Edition below says what it shares.
When a cloud transcription succeeds, SolScribe generates a downloadable Certificate of Destruction, but only after it confirms the staged audio was deleted or that none was staged. The certificate lists the job IDs, audio duration, estimated file size, processing tier and a SHA-256 hash of the job metadata rather than of the audio. It states where the staged copy was deleted from and the UTC time that was recorded, or that no copy was staged.
A certificate attests only to what SolScribe observed, not to copies it cannot see, such as the GPU worker's memory. Failed and cancelled jobs get none. If the delete at completion can't be confirmed, the fallback sweep issues the certificate after its own delete, stamped with the sweep's time. A recording uploaded in pieces and merged gives one download holding each piece's certificate, and it names any piece without one. Audio destruction is also written to the immutable audit log (actor, action, record ID, timestamp), which HIPAA-tier accounts can export as CSV.
| Data | Why | Where | Retention |
|---|---|---|---|
| Audio files | Transcription only | A RunPod GPU worker; uploads over about 5 MB are staged in AWS S3 (SSE-KMS) first | Staged copy deleted from S3 when the job completes; a fallback sweep retries a failed delete |
| Transcripts | The product you paid for | Neon Postgres, on storage encrypted at rest (AES-256) | Until you delete them |
| Account data (email, password hash) | Authentication | Neon Postgres + Neon Auth | Life of account |
| Compute ledger (minutes used) | Bill the correct number of minutes | Database | Life of account; no audio content or filenames logged |
| Payment data | Process payments | Stripe (we never see card numbers) | Per Stripe |
| Operational logs / errors | Reliability and debugging | Sentry (errors) + Fly.io (server logs) | Sentry error events retained up to ~90 days; operational data only — verified to carry no transcript, audio or PHI (user reduced to country-level geo). Fly.io server logs are ephemeral. |
All data is transmitted over TLS. Transcripts are stored in Neon Postgres, whose storage is encrypted at rest with AES-256. Staged audio is encrypted at rest in AWS S3 with KMS-managed keys until it is destroyed.
Your audio and transcripts are never used to train AI models — ours or anyone else's. They are processed solely to produce and store your transcript.
SolScribe offers optional features such as chat-with-your-audio and summarization. For the hosted service this inference runs over your transcript text only — never audio — via Anthropic, on our account. You can instead use your own OpenAI or Anthropic key (Bring-Your-Own-Key) by choosing that provider and saving its key through the SolScribe API, since the web app has no AI settings page. AI features for your account then send your transcript text to that provider on your account, under your agreement with them, and a request that fails on your key is not retried on ours. A maintenance pass we run over stored transcripts, such as re-checking their quality or extracting knowledge from older ones, uses our account whatever your settings.
We rely on the following service providers to operate SolScribe:
The marketing site uses self-hosted Plausible analytics (no cookies, no personal data). The hosted application itself contains no analytics or advertising trackers.
SolScribe's self-hosted edition runs on your own hardware. From version 1.8.3.7 it sends nothing to us: no audio, transcripts, analytics or license checks. Your audio and transcripts stay on your server until you delete them.
Two components it bundles contact third parties by default. Hugging Face hosts the speech models, so it sees your server's IP address and which models the server downloads or checks for updates. pyannote, the voice-detection library, reports each run to pyannoteAI with the recording's length, speaker settings, library version and a random session ID. pyannoteAI sees your server's IP address but never the audio or text. Set PYANNOTE_METRICS_ENABLED=false in the server's environment to stop those reports.
Anything else leaves your server only if you set it up, and goes straight to that service rather than to us. If you connect OpenAI or Anthropic with your own key, SolScribe sends that provider your transcript text, including for the titles and summaries it writes after each transcription. A webhook sends transcripts to the address you give it, and cloud transcription through your own RunPod account sends audio to RunPod.
We're not offering HIPAA processing yet. The managed HIPAA tier is on hold, and DST Digital LLC is not signing Business Associate Agreements (BAAs) with customers while it is. Do not send Protected Health Information (PHI) to the SolScribe cloud service. You can join the waitlist to hear if that changes.
SolScribe is operated by DST Digital LLC. Privacy questions or data requests: privacy@solscribe.ai / compliance@solscribe.ai.
We will update this policy as the service evolves and bump the "Last updated" date above.