Profile
API keys
Keys for the Olewave APIs (Olign, and the Voice Clone SDK): send one as the
X-API-Key header. Make one per server, so each can be replaced on its own: create the new key,
deploy it, then revoke the old one. A key is shown once, when it is created.
Terms
Olign
APIForced alignment and pronunciation scoring of recorded English speech: word and phone timings with scores. 3 credits per minute of audio during the launch promotion, then 3.5, billed per second (at least 10 s a call); see Billing.
Voice Clone
by invitationTrain a custom voice from a speaker's recordings.
Olign quick start
Upload a recording, optionally with its transcript, then poll until it is done. Your key is under Profile.
curl -H "X-API-Key: $OLIGN_KEY" -F [email protected] -F [email protected] \ https://api.olewave.com/olign/v1 # {"job_id":"6abafb11aac446cc"}
curl -H "X-API-Key: $OLIGN_KEY" https://api.olewave.com/olign/v1/6abafb11aac446cc
Up to 100 MB and 3 hours of audio per upload. Each minute of audio
costs 3 credits during the launch promotion and 3.5 after it, billed per second (rounded up, at least 10 s), charged when the job is
accepted and refunded if it fails; without enough credits the
API answers 402 (see Billing). Compress long recordings (Opus at 32 kbit/s is about 14 MB an hour):
ffmpeg -i in.wav -ac 1 -c:a libopus -b:a 32k out.ogg. Your audio is deleted as soon as the
request finishes, and your transcript and result as soon as you fetch the result (24 hours after the
request if you never do); nothing you send is used for training; usage
is counted by audio duration. Terms.
Add a speaker / upload more data
A speaker is a voice you want to clone. Give us a few minutes of that person's clean speech and we'll train a voice model from it. Two ways to add it:
Upload recordings
easiestPick audio files from your computer, or zip the whole folder and drop that in — right-click → Compress on Mac, Send to → Compressed folder on Windows. They're stored privately in your own space; no one else can see them.
WAV · FLAC · MP3 · M4A · OGG · OPUS, or a .zip of them — a few minutes total is plenty.
Transcripts are optional: put a clip1.txt next to clip1.wav (or a transcripts.csv of filename,text in the zip) and we’ll use your text; otherwise we transcribe it for you.
Push from your server
SDKHave a backend with lots of recordings? Our SDK examines each file locally (duration, silence, clipping) and pushes only the clean ones — straight from your server, with your server API key.
Uses your API key — create it under Profile.
🔒 Server-side only. Run this on your backend — never in a browser, mobile, or desktop app. The key is a secret; anyone who has it can push data as you. Rotate it anytime.
Your speakers
Your models — bundle voices into one deployable TTS model
pick one or more voices
Train a voice
Training history
Audio per day
Daily numbers
By operation
By API key
Recent calls and jobs
Credits
Recent activity
Usage limits
A monthly limit stops Olign once this month's usage reaches it: calls are
answered 429 with Retry-After until the 1st (UTC), and nothing more is charged.
Leave it empty for no limit.
Password
Storage — bring-your-own S3
Store your trained deltas (and read job I/O) in your own bucket. Saved server-side; leave blank to use Tycho's default storage.
To delete your account and the data kept for it, write to [email protected].