We proudly offer

Olign — Olewave’s Lancet-Accurate Speech-To-Text Forced Alignment Service

Send a recording with its transcript, or with an ASR result. Without either, Olign transcribes the audio itself. Every word and phone comes back with its start and end time and a score from 0 to 100.

Sign up, add a card and get 3,000 free credits, 1,000 minutes of audio at our launch price. Saving the card charges nothing. Olign is in beta, and signing up takes an invitation code, which you can request from us.

Word timestamps are often less accurate than they look. The FA-Bench paper measured how far off they are. Whisper places words about 150 ms early. Seven of the eight commercial speech-to-text APIs start their words late, three of them by more than 50 ms. Forced aligners fail differently. With background noise added to the Buckeye test split, Montreal Forced Aligner 3.4 returned no alignment for 354 of the 4,513 utterances.

We built Olign to place word boundaries accurately on real recordings, including noisy ones.

Word boundaries scored against human annotators

FA-Bench hands every forced aligner the same transcript and scores each word boundary against the one a human annotator placed. On TIMIT core-test its word boundary F1 is 0.782, against 0.644 for Montreal Forced Aligner 3.4. F1 is the benchmark's primary metric because it counts every word a system fails to place. Mean absolute error leaves those words out.

Timestamps without a transcript

Send no transcript and Olign transcribes the audio itself. From audio alone, its word boundary F1 on TIMIT core-test is 0.756, against 0.443 for Google Chirp 2, the highest of the eight other commercial speech-to-text APIs in the benchmark.

Alignments that hold up on noisy audio

Across the Buckeye test split, clean and under four kinds of degradation, Olign returned an alignment for every one of the 22,565 utterances. Montreal Forced Aligner 3.4 returned nothing for 573 of them.

Numbers you can check

Every result on this page comes from FA-Bench. Olewave makes both Olign and FA-Bench. Olign's numbers are from Olign 1.0 in the October 2026 records, and the FA-Bench Leaderboard shows every system's. FA-Bench is public on GitHub and described in a paper on arXiv. The paper reports Olign 0.9, whose API stays available, so you can reproduce its numbers and run the same benchmark against your current stack.

Self-hosted when audio cannot leave your network

If your recordings have to stay inside your own VPC, we can discuss a self-hosted deployment.

OlignMFAQwen3-FAWhisperX

Word boundary F1 at 20 ms, and higher is better. A boundary counts as correct only when the words on both sides of it match the reference and it lands within 20 ms of the hand-labelled boundary. The splits are TIMIT core-test and Buckeye test from the October 2026 snapshot of FA-Bench. Scoring uses its fabench protocol on a 10 ms grid, with bootstrap 95 % confidence intervals. All four systems run in the gold-transcript track, so each is handed the reference words and only its timing is measured. No recognition error enters these numbers. Both splits are held out and speaker-disjoint. TIMIT core-test excludes the SA sentences, as the corpus documentation requires. The Buckeye test split holds 8 speakers, two from each of the corpus's four sex and age groups, and its utterance list is committed in the repo. Noisy is the mean of the four degradations, which are reverb, noise, music and babble. Baselines are Montreal Forced Aligner 3.4.1, Qwen3-ForcedAligner-0.6B and WhisperX 3.8.6. Full results are on the FA-Bench Leaderboard.

Olign is available now for US and UK English, with more languages planned. A minute of audio costs 3.5 credits, and 1,000 credits cost $1, so an hour of audio costs $0.21. During our launch promotion a minute costs 3 credits, so an hour costs $0.18. New accounts get 3,000 free credits when they add a card. You send each recording as one job, up to 3 hours long, and an hour of audio finishes in about four minutes. The accuracy figures on this page are measured on American English.

Start with 3,000 free credits.

That is 1,000 minutes of audio at our launch price, enough to measure Olign against the aligner or ASR you use today, on your own recordings. Adding a card unlocks the credits and charges nothing. If you don't have an invitation code yet, ask us for one.