Privacy Policy

Clea is a service operated by Gumburtu AI, based in Leeds, United Kingdom. Gumburtu AI is the data controller for the personal data described on this page — if you want to exercise a data right, that is the entity to address, at hello@clea.app.

• Account details. Your email address, and your name and avatar if you sign in with Google.
• Media you upload. The audio and video files you submit, and any media we fetch from a URL you paste.
• Transcripts and derived content. The transcript, speaker labels, AI summaries and insights, chat history for a project, and documents you export.
• Usage data. Minutes transcribed and number of projects, used to apply plan limits.

Uploads travel over TLS and are stored in private buckets that are not publicly listable. Media is read only through short-lived signed URLs generated per request.

Access to your projects is enforced at the database level using row-level security, so a request authenticated as another user cannot read your rows even if the application layer had a bug.

Clea is not able to transcribe or analyse recordings by itself. To provide the service, the following processors receive data:

• Supabase — hosts the database, authentication and file storage. It holds your account details, media, transcripts and derived content.
• AssemblyAI — performs speech-to-text and speaker diarization. It receives a time-limited signed URL to your media file for the duration of the job.
• Anthropic — generates summaries, insights, chapter titles, translations and answers to questions you ask about a recording. It receives the transcript text, not the audio.
• Vercel — hosts the application and processes requests.

Each of these providers has its own terms governing retention and training use. If you are handling recordings subject to specific regulatory duties, review those terms before uploading.

Files and transcripts are kept until you delete them. Deleting a project deletes its media, transcript, speaker labels, insights, chat history and generated documents.

You can also enable automatic deletion of source media once transcription completes, which keeps the transcript while removing the original recording.

Deleting your account removes all of the above. Backups are rotated on a rolling basis and purge deleted content within 30 days.

Depending on where you live you may have the right to access, correct, export or erase your personal data, and to object to certain processing. Every project can be exported from the application at any time in several formats. For anything the interface does not cover, email hello@clea.app.

• We do not sell your data.
• We do not use your recordings or transcripts to train our own models.
• We do not make your projects public unless you explicitly create a share link.

Questions about this policy: hello@clea.app.