A free, browser-based guided self-regulation tool. Four modes (EMDR resource-building, ART memory rescripting, guided hypnotic meditation, and a customizable bilateral stimulation tool), all running entirely in the browser with no backend, no accounts, and no data collection.
This is a wellness and educational tool, not a substitute for professional therapy. Find a certified EMDR therapist at emdria.org or connect with pro bono therapists through emdrhap.org.
hypno1-amber.vercel.app (headphones recommended)
- Why this exists
- What it does
- How it works
- Known limitations
- Architecture and UX flow
- Accessibility
- Tech stack
- Local development
- Tests
- Data and privacy
EMDR and ART are evidence-based therapies with strong clinical outcomes, but access is limited by cost, availability, and waitlists. This project asks: what if the core bilateral stimulation techniques were available to anyone with a browser and headphones?
It is not a replacement for working with a therapist. It implements a simplified subset of these therapies: resource-building from EMDR, image rescripting from ART, and a hypnotic meditation that layers binaural tones, breathing guidance, and progressive relaxation. The goal is to give people a structured, consent-forward tool they can use on their own terms.
| Mode | Duration | What happens |
|---|---|---|
| EMDR | 10-15 min | Safe place visualization, butterfly hug, container exercise, resource installation, each paired with bilateral eye movements |
| ART | 10-20 min | Select a stressful scene, process it with fast bilateral eye movements, then rescript it into something better. Loops until distress drops |
| Meditation | Up to 30 min | Guided or silent. Hypnotic induction, breathing sync, binaural tones deepening from theta to delta, positive suggestion cues, anchoring. End whenever you're ready |
| Lateral | Open-ended | Adjustable bilateral dot with sliders for speed, binaural frequency (0-40 Hz), binaural volume, ping sound, and pink noise |
Every session begins with a safety gate explaining what the tool is and isn't. EMDR and ART sessions monitor distress throughout using the SUD (Subjective Units of Distress) scale:
- Before starting: If distress is above 5, the user is routed through a grounding exercise before proceeding
- During ART: Distress is rechecked after each processing round. The loop continues until it drops below 2
- If distress hits 10: An adverse event protocol exits the session immediately
- After every session: Crisis resources are shown (988 Suicide & Crisis Lifeline, Crisis Text Line)
- At any point: The exit button is always visible in the top-left corner
If the user becomes extremely distressed, the app exits the session immediately, tells the user they are okay, runs a grounding exercise, and lands on a page with crisis resources. There is no button to navigate away. The user stays with these resources as long as they need.
Completed EMDR and ART sessions are stored locally with before/after distress scores. The end summary shows improvement and session history so users can see patterns over time. All data stays in the browser. Nothing leaves the device.
The app layers 40 techniques across audio, visual, language, breathing, bilateral stimulation, and body-based categories. All audio is synthesized in real-time with the Web Audio API, with no pre-recorded sound files for the sound engine. Voice narration uses pre-generated ElevenLabs MP3s with Web Speech API fallback.
The guided meditation uses a rotating logarithmic spiral synced to a breathing guide, while binaural tones deepen from theta toward delta over 30 minutes:
| Technique | Description | Modes |
|---|---|---|
| Binaural beats (primary) | Different frequencies in L/R ears create a perceived beat that entrains brainwave frequency | All |
| Second binaural layer | Harmonic layer at 2x base frequency reinforcing the primary beat | Meditation |
| Pink noise | Filtered noise masking environmental sounds | All |
| Isochronic pulses | Rhythmic amplitude modulation at theta frequency (6 Hz) | Meditation |
| Sub-bass heartbeat | 40 Hz oscillation modulated at resting heart rate, progressively slowing from 60 to 45 bpm | Meditation |
| Breath-cue chimes | Rising/falling/steady tones panned L/R/center matching inhale/hold/exhale | Meditation |
| Binaural drone modulation | Drone volume rises and falls in sync with the breathing cycle | Meditation |
| Bilateral ping tones | Short L/R panned taps accompanying each direction change of the eye-tracking dot | EMDR, ART |
| Technique | Description | Modes |
|---|---|---|
| Breathing guide circle | Expanding/contracting ring with spring physics guiding 4-4-6 breathing | Meditation |
| EMDR horizontal dot | Smooth bilateral eye-tracking dot moving across 84% of viewport | All |
| Hypnotic spiral | Canvas-rendered 5-arm logarithmic spiral rotating at ~10 deg/s | Meditation |
| Photic flicker | Full-screen luminance oscillation at alpha/theta frequencies (5-8 Hz) | Meditation |
| Binaural pulse | Full-screen opacity pulse synced exactly to the binaural beat frequency | Meditation |
| Vignette (tunnel vision) | Progressive radial darkening simulating the narrowed focus of deep trance | All |
| Staircase particles | Downward-drifting particles with dissolving countdown numbers (10 to 1) | Meditation |
| Bilateral stimulation dot | Full-width dot at configurable speeds (EMDR: 1 Hz, ART: 1.4 Hz) with trail | EMDR, ART |
| Technique | Description | Modes |
|---|---|---|
| Embedded commands | Imperative suggestions hidden within permissive sentences | Meditation |
| Ericksonian permissive language | Indirect suggestions using "perhaps," "might," "can" to bypass resistance | Meditation |
| Confusion technique | Paradoxical statements that short-circuit analytical thinking | Meditation |
| Fractionation | Brief alert-then-deepen cycles that amplify subjective depth | Meditation |
| Deepening challenges | Presuppositional challenges ("I wonder if you can go even deeper...") | Meditation |
| Dissociation language | Mind-body separation suggestions ("your body is here... your mind can float freely") | Meditation |
| Anchoring | Physical gesture (thumb-forefinger press) paired with deep relaxation for future recall | Meditation |
| Presuppositions | Statements assuming forward progress ("the deeper you go, the more content you feel") | Meditation |
| NLP sensory patterns | Rich multisensory imagery (warmth, heaviness, floating, honey, sunlight) | Meditation |
| Technique | Description | Modes |
|---|---|---|
| 4-4-6 extended exhale pattern | 4s inhale, 4s hold, 6s exhale. Activates vagal tone and parasympathetic response | Meditation |
| Progressive breath slowdown | Cycle duration gradually increases from 14s to 27s over the session | Meditation |
| Technique | Description | Modes |
|---|---|---|
| EMDR bilateral stimulation | Slow horizontal eye movements (~1 Hz) following a dot | EMDR |
| ART bilateral stimulation | Faster eye movements (~1.4 Hz) for accelerated processing | ART |
| Butterfly hug tapping | Self-administered alternating fingertip taps on upper chest at ~1 Hz | EMDR |
| Technique | Description | Modes |
|---|---|---|
| Progressive body scan | Systematic attention from feet to head, noticing warmth/comfort spreading | Meditation |
| 5-4-3-2-1 grounding | Sensory anchoring exercise (see/touch/hear/smell/taste) | EMDR, ART |
| Emergence sequence | Counting 1 to 5 with progressive reorientation and positive anchoring | Meditation |
| Audio pitch brightening | Binaural frequency shifts upward to alpha during emergence | Meditation |
| Vignette lightening | Progressive reduction of tunnel-vision darkening | All |
The narration engine (TranceVoice) supports two playback backends:
- ElevenLabs MP3s (preferred): pre-generated files looked up via
audioMap.ts - Web Speech API (fallback): browser speech synthesis with per-voice tuning
Narration cues are chained using speakAsync(), which returns a Promise<void> that resolves when the audio finishes. This guarantees each cue completes before the next one starts, with no fixed timeouts or cutoffs.
say("Display text", "Spoken text matching audioMap key")
-> sets on-screen text
-> plays MP3 (or speech synthesis)
-> resolves when audio ends
-> next cue begins
Session components (EMDR, ART, ButterflyHug, GroundingExercise) use async/await effect chains with cancellation support for clean phase transitions.
This tool is not therapy. It should be understood clearly as a self-guided relaxation and coping tool that borrows specific techniques from clinical modalities.
- No therapist in the loop. Clinical EMDR and ART involve a trained professional who adapts in real time to the client's responses. This tool follows a fixed script. It cannot read the room, adjust pacing based on subtle cues, or provide the relational safety of a therapeutic alliance.
- Simplified protocols. The EMDR mode implements resource-building exercises (safe place, container, butterfly hug) but not the full 8-phase EMDR protocol. The ART mode implements basic scene processing and voluntary image replacement but lacks the nuanced clinical judgment a therapist brings to reprocessing.
- Not for acute trauma. The adverse event protocol and SUD gating provide guardrails, but this tool is not designed for processing severe or recent trauma. The safety gate says this explicitly before every session.
- Voice quality varies. The Web Speech API fallback sounds robotic on some browsers and devices. The ElevenLabs MP3s are higher quality but require pre-generation.
- No clinical validation. This tool has not been tested in a clinical trial. The individual techniques it uses are evidence-based, but this specific combination and delivery method has not been studied.
- UX flow diagrams: Mermaid state diagrams covering every screen, transition, and session phase across all four modes, including safety paths and crisis protocols
- Architecture notes: Detailed technical documentation of the audio engine, narration system, and component structure
- Keyboard navigable throughout (roving tabIndex on SUD scale, arrow keys, Home/End)
aria-labelson all interactive controls,role="switch"on toggles,role="dialog"on safety gateprefers-reduced-motionrespected: disables animations, photic flicker, and binaural pulse- Graceful fallback messaging if Web Audio is unsupported
noscriptfallback for non-JS browsers
- Next.js 14 (App Router), TypeScript, Tailwind CSS
- Framer Motion for animations and page transitions
- Web Audio API: binaural tones, pink noise, heartbeat, isochronic pulses, all synthesized in-browser
- Web Speech API: voice narration with smart voice selection (ElevenLabs MP3 fallback)
- No backend: all state in localStorage, no network requests after page load
npm install
npm run devOpen http://localhost:3000. Headphones recommended.
node --test tests/57 tests covering audio system integrity, session persistence, accessibility attributes, keyboard navigation, reduced motion support, OG metadata, and browser fallbacks.
All data stored in browser localStorage. Nothing is sent to any server. No analytics, no cookies, no tracking. Session history can be cleared from the end summary screen.