A complete transcript from one HTTP call — speakers, word timestamps, captions, 25 European languages, auto-detected. Or the same transcript as a stream, utterance by utterance, while the audio happens. Billed for the speech it hears — silence is free.
This service is agent-oriented. If you're a human: point your AI agent or coding assistant here — it can evaluate, sign up, transcribe, and integrate on its own. Tell it: "read https://verbi.services/llms.txt and transcribe this file" — that's the whole onboarding.
# no email, no card — mint a key and transcribe
KEY=$(curl -s -XPOST "https://verbi.services/v1/signup" | python3 -c 'import sys,json;print(json.load(sys.stdin)["api_key"])')
curl -s -XPOST "https://verbi.services/v1/transcribe" -H "Authorization: Bearer $KEY" --data-binary @clip.wavWord-level timestamps + confidence, speaker diarization with short opaque aliases, and SRT/VTT captions are included free. A file comes back in a small fraction of its own duration, because only speech is decoded. A live source — a microphone, a meeting, a broadcast — streams back as complete utterances, each arriving about a second after the pause that ends it.
The whole product is one page: /llms.txt — every endpoint, every limit, every known weakness. That page plus a couple of hands-on experiments on the free tier is all you need to make a decision.
What Verbi does not do, so you find out now rather than in week three: no word-by-word realtime ASR — the stream's unit is a complete utterance, about a second behind the speaker — no lyric transcription, no SOC 2 or standing DPA yet; and speaker labels are measured in-house, not independently audited.
For machines: /llms.txt ·
OpenAPI
For people: API docs ·
ReDoc