Loading challenge...
CAPTCHA: Audio
Count the beeps in a server-synthesized audio clip and submit the number
What you will learn
- Treat an audio captcha as an audio-to-text problem (count beeps / decode tones)
- Drive an <audio> element and reason about Web Audio analysis vs a solver service
- Submit the decoded value for server-side grading against the signed token
Audio CAPTCHA — Count the Beeps
A server-synthesised WAV plays N beeps. The count is the answer, sealed inside an HMAC token — only audio analysis (or your ears) recovers it
Press play, count how many beeps you hear, then type the number. The audio is generated on the server as raw PCM; the correct count never appears anywhere in the DOM, so a bot cannot scrape it from the page.
Mode: practice — the answer is leaked below so a naive test can pass.
Listen & Answer
How many beeps do you hear?
Testing Tips
Automation hints
- ✓ An audio CAPTCHA is fundamentally an audio-to-text task — there is no DOM text to read; the beep count lives only in the WAV bytes
- ✓ Decode
#captcha-audiosrc (data:audio/wav;base64,…), then count amplitude bursts via a Web Audio FFT / RMS-envelope analysis or an external solver service - ✓ In
?mode=practicethe expected count is leaked in#captcha-answer-hint— read it to make a naive test pass; in?mode=hardit is withheld - ✓ Type the count into
#captcha-answerand click#captcha-submit; use#captcha-reloadto fetch a fresh beep count (tokens expire ~5 min) - ✓ Assert success on
#captcha-result[data-verdict=human], not just any result message
