YouTube

Video is data nobody searches. A demo recording that shows a live dashboard, a webinar where someone reads out an account number, a training video with a staging environment on screen — all invisible until the words and the pixels are turned into text.

What you need to connect

Either channel URLs or handles (@YourCompany), or explicit video URLs — at least one of the two. Public videos need no credentials. For age-restricted or members-only videos, supply a cookie file.

What Classifyre reads

One asset per video, whose content is a transcript:

1. If the video has captions, they are used — in your preferred language order.
2. If it doesn’t, the audio is transcribed by speech recognition.
3. On-screen text is read from sampled frames, so a password visible in a screen recording is found even when nobody says it out loud.

Video metadata — title, description, channel, upload date, view count — is recorded alongside.

Metadata on every asset

Lineage

Worth knowing

• Transcription costs time. A channel of hour-long videos is a long scan; the sampling strategy is what keeps each run bounded.
• Unlisted videos are scannable if you have the URL — and are exactly where internal recordings tend to live.
• A proxy can be configured when scanning at a scale that attracts rate limiting.

Configuration

Beyond the fields below, every source also has the settings shared by all of them: the sampling strategy, the detectors to run, the scan schedule, and the compute limits for its scan jobs.

Required

Without these, the source will not save.

Secrets

Stored encrypted and never shown again after you save them. See Configuration & Fields.

Optional

Everything you can tune. Sensible defaults apply when you leave them alone.