Privacy Policy

Last updated: 10 October 2026

Who operates CaptionFetch

CaptionFetch is operated by TritaLabs Pvt. Ltd. Contact contact@tritalabs.com for privacy, support, or security questions.

Information we process

We process submitted video URLs and caption text to provide transcripts. Accounts include an email address, authentication identifiers, preferences, saved transcripts and collections, query history, and payment references. We do not request YouTube passwords or store payment-card details.

Uploads and saved content

Uploaded audio and video are held in private storage and processed for transcription. Completed media may be retained to support playback and subtitle editing. Transcripts, custom vocabulary, brand preferences, generated drafts, source snapshots, verified excerpts, and draft revision history remain associated with your account until deleted. Access to media uses time-limited links; anyone you share such a link with may access it until it expires.

AI and translation

When you request AI output, the selected transcript, your question, and any saved brand preferences or revision instructions are sent to the configured provider: OpenAI or Moonshot AI (Kimi). Translation may use YouTube, Google Cloud Translation, or the configured AI provider. Audio transcription may use Groq, OpenAI, or Deepgram when speaker identification is enabled. These providers process content under their applicable terms and retention settings; we do not make an unconditional promise that providers retain no data. Only submit content you are authorized to share with these processors.

Source review and meaning search

Source corrections, preserved originals, saved moments, glossary terms, reviewed translations, background job progress, and prepared search passages are stored with your account. Preparing meaning search sends the selected transcript passages to the configured embedding provider, OpenAI by default; each meaning search sends your query to that provider. The resulting numerical representations are stored privately to compare meaning within your library. Deleting a saved source removes its related review and search records; saved draft snapshots remain part of that draft until it or the account is deleted.

Invited review

When shared review is available and you publish a saved source or draft for review, we store that text as a separate published snapshot. Invitations store the recipient email, role and acceptance status; the invitation credential is stored as a hash. Only the intended recipient using a verified account can accept it. Viewers can read and download published snapshots; reviewers can also comment. Private source notes, other library records and media files are not included. Later saved edits remain private until you publish them. Invitation links are kept temporarily in your browser session to return through sign-in. We do not send invitation emails automatically.

Comments retain the published version and paragraph or timestamp they refer to. Revocation stops subsequent access but cannot recall downloads already made. Account exports include accessible published reviews and your own comments, including your contributions to revoked reviews. Deleting the original shared resource or the account of its owner removes its shared snapshots, invitations and comments. Deleting a recipient account removes its invitations and authored comments.

Browser capture

The optional browser capture extension reads the active YouTube page URL only when you click its toolbar action. It opens the web workspace with that source selected, where existing login, limits and availability checks apply. It does not collect browsing history or read account credentials. The web app processes the submitted URL when you choose to fetch captions.

Security and service delivery

We use a salted identifier derived from network information for daily allowances and abuse prevention. The usage table does not contain the original IP address. Hosting and network providers may separately process connection information for security and delivery. Operational application logs omit transcript content and arbitrary provider error details. Provider accounting records operation identifiers, provider/model names, usage units and cost estimates or reconciled costs, including failed attempts. Billing events are kept independently of optional analytics.

Service providers

Service providers include Google Cloud, Supabase, Cloudflare where configured, Razorpay, Stripe, and the AI providers described above. Processing may occur outside India depending on the configured provider and infrastructure. We do not sell account data or transcript history.

Analytics and Google advertising

Optional analytics loads only after you accept it through the Cookies control. When advertising is enabled, we use Google AdSense on eligible public guide articles. Our transcript tool, user transcript results, dashboard, and account and legal pages do not have advertising placements. Google and other advertising partners may collect network and device information, including IP addresses, browser details, identifiers, and interactions with advertisements. They may use cookies, web beacons, or similar technologies to serve, personalize, and measure ads according to your consent and applicable settings. We do not include your saved transcript text or account email in advertising requests.

Google’s Privacy & messaging consent service loads when ads are enabled to obtain advertising choices, including a certified consent message for visitors in the EEA, UK, and Switzerland. Display-ad requests wait for our local opt-in and the consent service’s applicable decisions. Rejecting optional storage does not prevent you from using the free tool, signing in, or paying for a plan. Paid plans exclude our advertising placement.

Use the persistent Cookies control to change your local choices and Advertising privacy settings, when available, to reopen Google’s consent message. Google may also provide a Do not sell or share link for applicable US state privacy choices. Learn how Google uses information from sites that use its services and manage Google ad personalization through My Ad Center.

Retention and your choices

Anonymous transcripts are returned to the browser without being saved to an account. Registered users can remove saved transcripts and history, export account data, or delete their account from the dashboard. Complete active processing and cancel recurring subscriptions before deleting an account. Financial references may remain with us or payment providers where accounting obligations require retention.

Requests and concerns

For access, correction, deletion, consent withdrawal, or a privacy concern, email contact@tritalabs.com. We may need to verify account ownership before fulfilling a request.