Audit desk
From transcript to audit dossier
Paste the transcript, watch the cost update live and decide whether the title pack comes now or later. No token is spent before you start.
Put your video under audit
Paste your video transcript and let your AI auditor go through it line by line. (Any language.)
200 more characters to unlock the audit
How it works
Three steps, one honest report
- 01
Paste the transcript
Any language works. From 200 to 300,000 characters; every started 5,000 characters is 1 token.
- 02
The auditor reads every line
It scores 12 watchability metrics and checks the script against YouTube's Community Guidelines and advertiser-friendly rules, quoting the exact lines it flags.
- 03
Fix, export, publish
Apply the pinpoint fixes and safer rewrites, grab the PDF, or send the report to any AI with a ready prompt that applies every fix.
Inside the dossier
Every part of the report has a job
Read it top to bottom or jump straight to the part you need. Scores are colored by the same thresholds the auditor scores with.
- 01
Auditor's summary
The verdict in a few sentences, with the overall score.
- 02
Your strengths / Leak points
What already works, and where viewers are likely to drop off.
- 03
Retention map
How the script holds attention from the opening to the end.
- 04
Watchtime Report Card
12 cards, each with findings, risks and power moves, colored at 85 and 70.
- 05
YouTube Policy Radar
Flagged lines quoted verbatim, why they are risky, and safer rewrites in your transcript's language.
- 06
Final touches before publish
A short pre-publish checklist.
- 07
Pinpoint fixes
Spot, current line and an upgraded version you can paste.
- 08
Title & description+1 token
Optional: click-earning titles, description, tags and keywords for +1 token, now or later.
What the radar checks
Two rulebooks, fourteen categories
Strike-level rules and monetization rules carry different consequences, so the auditor keeps them apart and tells you which one a line touches.
Community Guidelines
Risk: strikes or removal
- Hate speech
- Harassment and demeaning content
- Harmful or dangerous acts
- Health misinformation
- Civic misinformation
- Child safety
- Graphic violence
- Regulated goods
- Spam and deceptive practices
Advertiser-friendly guidelines
Risk: limited or no ads
- Profanity
- Adult themes
- Controversial and sensitive topics
- Financial claims
- Dangerous acts
Risk levels
Low risk
Defensible in context, but a more cautious phrasing exists.
Medium risk
Likely limited ads or a borderline strike.
High risk
A plausible strike or removal.
Every evidence quote is checked against your transcript; a finding whose quote cannot be found there is removed before you see the report.
FAQ
Before you spend a token
Every started 5,000 characters counts as 1 token, so a 12,000-character transcript costs 3. The title and description pack adds 1 token, either right away or later from the report screen.
Any language. The report is written in the site language you have selected, while safer rewrites and pinpoint fixes stay in your transcript's language so you can paste them straight in.
You see an error and no tokens are charged; simply try again. If the title pack can't be produced, its token isn't charged either.
Each finding has to quote your transcript word for word. Quotes are checked against the text you pasted, and a finding without a matching quote is removed before you see the report.
Use Export to AI: choose the sections, add the apply-the-fixes prompt and hand the file to any AI. The prompt asks it for your fully revised transcript with a short change log.
Past audits keeps your 20 most recent audits with their scores, so you can reopen a report, view the script or export it again.
Your script, audited before it airs
Paste your video transcript and let your AI auditor go through it line by line. (Any language.)