Transcription your agent can buy by itself. Every word, timestamped.
Point WAVE at any recording and get back {text, words[]}. Every word carries its own timestamp, ready to search, seek, and caption. Pay by the call. No account, no subscription, no SDK to install.
One call, every word timestamped
curl -X POST https://api.wave.online/v1/transcribe \
-H "Authorization: Bearer <key>" \
-H "Content-Type: application/json" \
-d '{"source":"https://example.com/clip.mp3","lang":"en"}'
→ {"text":"...","words":[{"w":"...","t":0.4}, ...]}
Every word carries its own timestamp, so "search the transcript" and "jump to that moment" are the same click. No account or subscription in the way. Pay per call, settled through the WAVE gateway, the same bearer key and meter that run every other WAVE spoke.
Speech-to-text a machine can act on
Because it lives on the WAVE protocol plane, a transcript is a typed result. Your agents pay for it and pipe it straight into search, live captions, or the next WAVE service, all under one key and one meter. Speech-to-text becomes something machines buy and act on, not a subscription someone has to hold.
Federated. Auth, entitlement, and metering run through api.wave.online. This spoke renders the front-door and forwards the rest.