Content
What gets refused, why resending fails, and your duty to label AI ads.
Read as MarkdownWhat RiffAds refuses to make, and what you must do with what it makes.
One check on every path
REST, MCP, the CLI, the app and workflow runs all run the same check, so a request refused on one is refused on all. It runs before anything is charged, so a refusal carries credits_charged: 0.
What is checked
| Input | When |
|---|---|
| Script, prompt and other text fields | On submit |
| Uploaded or imported images | When the file is finalized, before use |
| Video and audio | Not scanned (content_checked: false) |
Text is matched against a fixed rule list, and a model pass can catch wording the list misses.
| Category | Result |
|---|---|
| Sexual content involving minors | Refused |
| Explicit sexual content | Refused |
| Credible violence and threats | Refused |
| Illicit goods | Refused |
| Hate | Refused |
| Self harm | Warning, job runs |
| Impersonation | Warning, job runs |
Video and audio are never reviewed, so don't tell your users they were.
What a refusal looks like
moderation_blocked(422, not retryable), with one sentence and no category.- A refused file names its
ast_id. That file can never be used. - An upload still being scanned returns
input_not_ready, which is retryable.
Resending fails
Same words, same refusal. 2 refusals from the rule list block the request for 30 days: Blocks.
- A blocked request returns
request_blocked. Change the script or the inputs. - If checks are down, submits return
moderation_unavailable(retryable) and reads still work. - There is no appeal button. A different request gets a fresh check.
Everything is AI made
Every file RiffAds returns is AI generated or AI edited, even with your own script or uploads. RiffAds adds no watermark, visible mark or C2PA data.
Labeling is your job
RiffAds never publishes anything, so the duty to disclose stays with you. No response tells your agent to label: build it into your process.
Label where law, a regulator or a platform requires it:
- EU AI Act: disclose realistic AI content (in force since 2 August 2026).
- Laws on deepfakes, synthetic media, political ads, and use of a person's name, image or voice.
- Advertising and consumer law, including fake reviews, testimonials and endorsements.
- The ad policies of every platform you run on.
Use a label people notice, like "AI generated", and burn it in with Text Overlay: add exact words. This is not legal advice.
Actors are not real people
- Library actors are AI. Never present one as a real customer, reviewer or expert.
- An actor from a real person's photo, or a cloned voice? Get their written consent first, and keep a copy.
- Never anyone under 18. Never a public figure without written consent.
- RiffAds does not check who is in a photo or voice sample, so the duty is yours.
Review before you publish
Nobody at RiffAds watches output, and the check runs on inputs, not on the finished file. Expect sometimes:
- Wrong, missing or mispronounced words.
- Lips out of sync. Odd faces, hands or text.
- False product details or claims, including in AI-written scripts.
- Accidental likeness to a real person or brand.
Full rules: AI Disclosure and Acceptable Use Policy.