Audio and video files you submit to Signal Loom AI for transcription are processed ephemerally. They exist in system memory only for the duration of the transcription job โ typically seconds to a few minutes depending on file length.
Once transcription is complete and the output has been returned to you, the original media file is deleted from our systems. We do not maintain a media archive, a content library, or any form of persistent storage for your uploads unless you explicitly opt into a storage feature.
Signal Loom AI does not use your prompts, uploaded media, transcripts, or API usage patterns to train, fine-tune, or improve any machine learning model โ including Whisper, any open-source model we run internally, or any third-party model we call.
This applies without exception to:
You retain full ownership of all content you submit to Signal Loom AI, including:
We claim no license, no implied license, and no rights to your content beyond what is strictly necessary to deliver the service to you.
When you opt into cloud transcription models (e.g., OpenAI Whisper API), your audio is sent to that provider's infrastructure. Those providers have their own privacy policies. We encourage you to review them:
Local transcription (MLX Whisper on Apple Silicon, or self-hosted) processes entirely within your own infrastructure or our cluster โ no media leaves your environment for those tiers.
For teams that deploy Signal Loom AI self-hosted or via private cloud, all media processing occurs within your designated environment. We do not receive, access, or log your media in those configurations.
Your API keys are hashed and stored securely. We cannot retrieve your raw API key after creation. If you believe an API key has been compromised, rotate it immediately from your dashboard. We do not have access to your active keys in plaintext.
| Commitment | Status |
|---|---|
| Media not stored after transcription | โ Confirmed |
| No model training on user data | โ Confirmed |
| No third-party data brokering | โ Confirmed |
| API keys hashed (not retrievable) | โ Confirmed |
| Data sold or traded to advertisers | โ Never |
| Prompts used for product improvement | โ Never |
| Media shared with third-party models without opt-in | โ Never |
For operational purposes, we retain anonymized, aggregated usage metadata:
This data is used solely to maintain service reliability, generate invoices, and detect abuse. It cannot be tied back to specific media files or transcripts.
You have full rights to your content. At any time, you can:
Questions about how your data is handled? We're here.
Email: privacy@signalloom.ai
Company: AIM Elemental Health Solutions, Inc.
Address: San Francisco, California, USA