AuraSTT
Open studio

Aura-3 speech engine

Transcription that keeps up with speech.

Upload or record audio and get word-level timestamps, speaker labels, and ready-to-ship subtitles in milliseconds.

Start transcribingTry the demo
<300ms
Edge turnaround
99.2%
Accuracy
Multi-speaker
Diarization built in
From $50
Pay-as-you-go credits

Live demo

Click a word, jump to the moment.

Interactive Web Audio Island

AI Architecture & Aura-3 Review

Click any word to jump audio timestamp. Word-level synchronization at 60 FPS.

0.0s/18.5s
Aura-3 Flagship
Alex (Lead Architect)Speaker 0

Aura-3demonstratedathirtypercentreductioninWER,especiallywhendealingwithdomain-specificvocabularyandacousticoverlaps.

Sarah (Product Lead)Speaker 1

AndbecausewestreamdirectlythroughCloudflareWorkersedgenodes,turnaroundlatencyissub-three-hundredmilliseconds.

Includes multi-speaker diarization, SRT subtitles, and credit balance ledger.
Open Full Studio

Features

Everything between audio and a finished transcript.

Aura-3 engine

Trained on 10M+ hours of multilingual audio, with a 30% lower word error rate than standard Whisper Cloud on technical jargon.

Word-synced player

Playback lights up each word as it is spoken. Click any word to jump to that exact timestamp.

Speaker diarization

Turns are split by speaker. Rename speakers and see talk-time percentages at a glance.

Subtitle export

Download SRT, WebVTT, plain text, or structured JSON for any editing workflow.

Edge-native

Runs on Cloudflare Workers with D1, so responses stay fast everywhere and there are no cold starts.

Pay as you go

Deposit credits and pay only for the minutes you transcribe. No subscription, credits never expire.

Benchmarks

Faster and more accurate than the usual options.

AuraSTTOpenAI WhisperGoogle SpeechAssemblyAI
Word error rate5.8%8.4%9.2%7.9%
Processing time<300ms3.5 - 6.0s2.0 - 4.5s2.5 - 5.0s
Word-synced playerBuilt inThird-partyAPI onlyAPI only
Speaker diarizationIncludedUnsupportedExtra costExtra cost
SRT / VTT exportOne clickManualManualManual
BillingPrepaid balanceCard lock-inCloud billingMonthly minimum

Pricing

Deposit once. Pay only for what you transcribe.

No subscription. Minutes are deducted from your balance as you use them, and credits never expire.

Starter Deposit

Minimum balance deposit to activate transcription with our Aura-3 Flagship foundation model.

$503,000 minutes · $0.016 / min
  • 3,000 minutes of transcription
  • Deducted as you use at $0.016/min
  • Aura-3 and Aura-2 model access
  • Interactive synced audio player
  • Multi-speaker diarization
  • SRT and VTT subtitle downloads
Deposit $50 to Start

Pro Growth Deposit

Most Popular

Best for video creators, podcast editors, and rapid research pipelines.

$1006,500 minutes · $0.015 / min
  • 6,500 minutes of transcription
  • Aura-3 flagship priority queue
  • AI Executive Summaries & Topics
  • Never-expiring credit balance
  • Cloudflare D1 transcription storage
  • Full REST API & Webhook access
Deposit $100

Scale Enterprise Deposit

Bulk audio processing credits for engineering teams and automated ingestion queues.

$25018,000 minutes · $0.013 / min
  • 18,000 minutes of transcription
  • Lowest unit cost ($0.013 / min)
  • Dedicated API token concurrency
  • Automated invoice & receipt portal
  • Priority Cloudflare Edge routing
  • Direct developer Discord support
Deposit $250

Start transcribing in a minute.

Record from your microphone or upload any audio file. Credits start at $50.

Open studio