Does Clipses train on my recordings or transcripts?
No. Clipses does not build or fine-tune AI models, and your content is never used as training data by us. Your videos, transcripts, clips and reels stay in your account and are only processed to do the work you asked for: transcription, search, tagging and rendering.
Which AI providers see my content, and do they train on it?
Three providers process your content, each for one job:
- Transcription: AssemblyAI turns your audio into a transcript. Clipses has opted out of AssemblyAI using customer content for model training.
- Search and suggested tags: when you ask a question in Chat or accept suggested tags, the relevant transcript excerpts go to a large language model through OpenRouter. Every request is sent with a no-data-collection setting, which routes only to providers that do not store prompts or train on them.
- Meeting recording: Clipses.ai Notetaker records Google Meet calls through Recall.ai. Clipses deletes the vendor's copy as soon as the recording has been imported into your project.
The full list of providers is in our Privacy Policy.
What about AI assistants I connect myself?
If you connect Claude, ChatGPT or another assistant to Clipses, the transcripts and quotes it reads go to that assistant under its own terms, in the same way as anything you paste into a chat with it. Check that assistant's data settings. See Can a connected AI assistant see all my projects?
Can I opt out?
There is nothing to opt out of on the Clipses side, because your content is not used for training. If you would rather a recording was not processed by AI at all, do not upload it. Deleting a video removes it from Clipses and from search.
Still stuck? Email hello@clipses.ai.