How Qirava Works
Qirava turns a public link or a file you upload (video or audio, including WhatsApp voice notes) into a ready-to-use .md file for your AI agents or knowledge tools. The process has four steps and, in most cases, takes just a few seconds.
- 1
Step 1: Paste the link
On the home page you'll find a field to paste the URL, or the option to upload a file. Paste a public video or audio link, or upload your file (video, audio, or WhatsApp voice notes), and press the button to start. No account or sign-in needed. Before you continue, make sure you have the right to use the content you'll process; Qirava does not verify rights ownership, and that responsibility falls on the user.
- 2
Step 2: Subtitle lookup
Qirava first looks for available subtitles, either manually added or auto-generated captions made accessible by the platform. This path is fast and requires no audio processing.
- 3
Step 3: Audio transcription (optional)
If there are no accessible subtitles and the STT engine is active, Qirava shows a confirmation panel with the video details before processing. You decide whether to proceed. Audio is temporarily downloaded, transcribed, and then deleted.
- 4
Step 4: Download the Markdown file
The result is a Markdown file with the full transcript, ready to download. The content of your transcript is not kept on our servers after it's returned to you. We only retain minimal, aggregated operational data, never the transcript text.
What data Qirava processes
- The URL you provide: to locate the video.
- Transcript in memory: temporary; not stored in any database.
- Temporary audio: only if STT; deleted after processing.
- IP hash: for rate limiting; not your raw IP.
- Technical logs: success/failure, platform, video ID; no transcript text.
What we don't do
- We don't keep the content of your transcripts on our servers after we return it to you in the same response.
- When transcribing from audio, the audio is processed in a temporary file that is deleted when finished.
- Our speech-to-text engine runs on our own infrastructure, so your content is not sent to third-party AI APIs.
- We do keep minimal, aggregated operational data (for example, per-domain counts and usage controls), never the text of your transcripts. Logging, cache, metrics and retention details are described in our Privacy Policy.
Current limits
- YouTube only (free tier).
- One video per request.
- Daily request limit per IP address.
- When audio processing is active: maximum video duration applies.
Authorized use
You must have the right to access and process the video you submit. Qirava is not designed to bypass access restrictions or infringe copyright.