Help & contact

A 5–10 minute audio script is usually finished in under 10 minutes, but longer scripts and library misses (when our visual library has no match for a scene) extend that. Watch the live progress page for an accurate ETA — it updates scene-by-scene as the pipeline runs.

We accept .mp3, .wav, .m4a, .aac, .ogg, and .webm— up to 200 MB. Scripts between 4 and 25 minutes give the best results today.

Failed renders show a trace_idon the project page. Include that id when you contact us — it lets us pull the exact pipeline log for your render. Common causes: audio with no detectable speech, scripts shorter than 60 seconds, or a transient upstream API outage (we automatically retry, but persistent outages can still surface).

We bill via Paddle. Pick a plan on the pricing page, complete checkout on Paddle’s hosted page, and your tier upgrades automatically when the webhook lands (usually within seconds). You can manage your card, download invoices, change plans, or cancel any time from the “Manage subscription” button on your account page.

Refunds are handled case by case in the first 14 days of a new subscription, especially for failed renders that we could not recover. Use the contact form below with category “Billing or subscription” and we’ll respond promptly.

Videos and intermediate artifacts are kept for 30 days, then automatically deleted. Download anything you want to archive permanently. See our privacy policy for the full data-retention summary.

Today we deliver a single rendered MP4 you can download. A scene-level editor is on the roadmap. If you want a specific scene re-rolled, the fastest path is to re-upload the same audio — the library will usually surface different candidates on the second run.

No. We use your audio only to generate the video you requested. We do not share, sell, or train models on your uploads. Full details are in our privacy policy and subprocessors list.