A hellish random tone generator. 15 oscillator voice types — sine, sawtooth, square, pulse, triangle, white noise, pink noise, FM synthesis, ring modulation, frequency sweeps, clicks, detuned stacks, glitch gates, and tremolo — fired in random combinations at random intervals. The devil watches over it all with glowing fiery eyes.
Random nonsense speech in 22 languages — English, Spanish, French, German, Italian, Portuguese, Russian, Japanese, Chinese, Korean, Arabic, Hindi, Turkish, Dutch, Swedish, Polish, Hebrew, Thai, Vietnamese, Indonesian, Greek, and Czech. Each word sounds intentional but means nothing. A desperate neighbor on a balcony with bloodshot red eyes watches the babel unfold. City sounds, formant-synthesized robot voice, and a podcast mode where two hosts argue in gibberish.
A submarine signal generator. Morse code beeps, shortwave radio static, electrical equipment hum, sonar pings with echo, dolphin clicks and whistles, owl hoots, ultra-low drones, whale calls, and ultra-random frequency sweeps from 2 to 20 seconds across the full audible range. A spy in a fedora and trench coat stands watch inside the submarine control room, surrounded by sonar screens and gauge panels. All sounds pre-generated in a buffer pool for glitch-free playback.
A real-time random oscillator engine that generates unpredictable tones using 15 different voice types. Each event has a randomly chosen frequency (20–7000 Hz, log-uniform distribution), duration (40–2000 ms), amplitude, and stereo pan. No two moments sound the same.
15 Voice Types: Sine, Sawtooth (2× oversampled with polyBLEP anti-aliasing), Square, Pulse (variable duty cycle via PeriodicWave), Triangle, White Noise, Pink Noise (Paul Kellet filter), Filtered Noise Burst (resonant bandpass), FM Synthesis (carrier + modulator with random ratio and index), Ring Modulation, Exponential Chirp, Percussive Click, Detuned Stack (3–5 beating sines), Glitch (gated at random rate), Tremolo Sine.
Generates phonotactically valid nonsense words in 22 languages using language-specific syllable inventories (onset consonants + vowels + codas). Each language has its own connector words for natural sentence flow. Speech is produced via the browser's SpeechSynthesis API with automatic voice discovery — the correct accent and prosody for each language comes free.
22 Languages: English, Spanish, French, German, Italian, Portuguese, Russian, Japanese, Chinese, Korean, Arabic, Hindi, Turkish, Dutch, Swedish, Polish, Hebrew, Thai, Vietnamese, Indonesian, Greek, Czech.
13 procedural signal generators inspired by submarine communications, marine life, and deep-sea ambience. All sounds are pre-generated into a buffer pool at initialization for glitch-free real-time playback — no main-thread blocking during performance.
All three oracles share the same professional-grade effects chain, built with the Web Audio API:
8 switchable ambient noise types, each with a per-type anti-alias filter:
Plus a persistent 60 Hz drone with 0.3 Hz LFO tremolo, independently controllable.