Last updated: October 10, 2026
Video and audio files are not kept long-term. Result files are kept for at least 24 hours. Uploads and scratch files are removed within a few hours. Saved project text is kept for 30 days on Free, 365 days on Pro, and 24 hours for a guest project that has not been moved to an account, once project saving is on for that account. Until then, transcript text is kept for at least 7 days. We do not use your content to train models.
When you upload a file, we process it on our infrastructure to deliver transcripts, subtitles, or other outputs. Video and audio are not stored long-term. Result files are kept for at least 24 hours. Transcript text from video, voice, and YouTube transcript jobs, and from Fix Transcript, is kept for at least 7 days, and never for less time than the result file, until project saving is turned on for the account. Once it is, saved transcripts, subtitles, translations, and edits are kept for 30 days on the Free plan and 365 days on Pro. A project started as a guest and not moved to an account is kept for 24 hours. Those periods are server configuration and can be longer; they are not shorter than the result file while that file still exists. Subtitle, translation, fix, convert, burn, and compress result files that are not saved into a project expire with the result file. Uploads and scratch files, including playback audio, are removed within a few hours (about two and a half hours, or about 30 minutes when the disk is more than 80 percent full). Job metadata, such as the job id and status, can remain after the file is gone. Connecting ChatGPT uses the same account and the same file retention. That connection stores the sign-in session (an access token for about 1 hour and a refresh token for about 30 days) and does not keep a separate copy of your files. Application logs do not keep a copy of uploaded files.
Uploaded video and audio for Video to Transcript, Voice to Text file uploads, and Video to Subtitles are sent to ElevenLabs for speech-to-text, including speaker labels when you turn those on. If ElevenLabs cannot complete a request, the same audio may be sent to OpenAI (Whisper) to finish it. Live microphone transcription uses Deepgram. Translation, summaries, and editing suggestions use OpenAI. Payments are processed by Stripe. These providers handle the content only to deliver the service you asked for. A legacy speaker-label provider (Replicate) is not used unless the server is explicitly configured to turn it back on. We do not use your content to train models.
If you sign up or subscribe, we store account information (e.g., email, plan) and usage data (e.g., uploads used today) necessary for billing and product limits. We do not use your content for training models or for any purpose other than providing the service you requested.
Sign up for a free account to use the tools. For paid users, you can manage your subscription and billing through our pricing page and the Stripe customer portal. We do not sell your personal information to third parties.
We may use essential cookies and analytics to operate the site and improve the product. We do not use your uploaded content for advertising or profiling.
VideoText, 11771 Miralago Blvd, Apt 1249, Dallas, TX 75234, United States. For privacy-related questions or requests, email [email protected] or use the Contact page.
This policy may be updated from time to time. The “Last updated” date at the top reflects the latest change.