ECHO
Vocal Looper & Scripting Sandbox — Echolalia Support
01. What It Is
ECHO is a private, fully offline loop station for two distinct but related autistic communication patterns: vocal stimming (repetitive vocalizations that regulate the nervous system) and scripting (rehearsing language for social situations in advance). Both practices are clinically legitimate self-regulation strategies, and both are currently underserved by technology.
The Core Problem It Solves
Many autistic individuals find hearing their own voice aversive — triggering the "cringe" response that makes script practice feel punishing rather than regulating. ECHO solves this by applying real-time pitch-shifting and heavy reverb via the Web Audio API, abstracting the voice into something ambient and non-personal. The result: rehearsal becomes regulating rather than dysregulating. The voice is present, but it no longer sounds "like me."
02. Features
Web Audio API Real-Time Vocal Processing
Microphone input runs through a live Web Audio node chain: pitch-shift (±12 semitones), reverb convolution (pre-loaded impulse response), and optional formant filter. No recording to disk. The audio exists only in the moment of playback through the headset. Privacy by architecture.
Loop Station Stim Loop Builder
Tap to record a loop (up to 8 seconds). The loop plays back continuously with pitch processing applied. Layer up to 4 loops simultaneously — each on its own pitch offset. The result is an ambient chord of your own voice, which many vocal stimmers find deeply regulating. Tap any loop to clear it.
IndexedDB Script Practice Cards
A simple card system for storing scripted phrases. "Phone call openers." "Grocery store interactions." "Explaining a diagnosis." Tap a card to see the text, record yourself saying it, hear it back through pitch processing. No social component — purely private rehearsal. Cards saved offline in IndexedDB.
Waveform Canvas Visual Feedback Layer
A gentle waveform visualizer renders the audio in real time — slow-moving, low-contrast, lavender-toned. Not a diagnostic tool. A visual stim companion to the audio experience, giving the voice a visual form without analyzing or critiquing it.