Turn a publicly reachable audio or video URL into text with per-segment timings and a detected language. Costs $0.05 per call, paid in USDC from your own wallet. The file must be under 800 KB; oversized files are refused for free before any payment.

Given subtitles or timings, return the spans of dead air and how much of the runtime cutting them would save. Costs $0.002 per call — a tenth of transcription, because the expensive half is not redone. Feed it the segments transcribe_url returns.

Free. Returns the prices, the size limit, the chain and how payment works, so an agent can decide whether to call the paid tools at all.