For people living with ALS (amyotrophic lateral sclerosis), voice banking is one of the most meaningful things they can do early: recording their voice while they still can, so that if speech is lost, their communication device speaks with their voice — not a stranger's. AI voice cloning has transformed what's possible, producing far more natural personal voices from far less recording than older technologies. But voice banking is time-sensitive, emotionally loaded, and full of practical questions. This guide covers voice banking for ALS in 2026: what it is, when to start, the approaches, the recording process, and how families can support it.
TL;DR: Voice banking means recording your voice now so a future communication device can speak as you. Start as early as possible after diagnosis — banking requires clear speech, and ALS progression is unpredictable. Modern AI cloning needs 30–60 minutes of good recordings (older programs need ~1 hour of scripted phrases). Record in a quiet room, follow the script prompts, involve family for support, and store the voice securely with your care team. It's an act of preserving identity, not just communication.
Key Takeaways
- Start voice banking as early as possible after diagnosis — it requires intelligible speech, and waiting risks losing the window.
- Modern AI voice cloning produces remarkably natural personal voices from 30–60 minutes of recordings.
- Older programs (like ModelTalker) need ~1 hour of specific scripted phrases and produce more robotic results — but they're free and proven.
- Record in a quiet space, in short sessions to avoid fatigue; consistency across sessions matters more than perfection.
- The banked voice integrates with AAC (augmentative and alternative communication) devices for daily use.
- This is emotional work — involve family, take breaks, and treat it as preserving identity.
What is voice banking?
Voice banking is the process of recording a person's speech to create a synthetic replica of their voice, stored for future use if natural speech is lost. For ALS — a progressive neurodegenerative disease that typically affects speech muscles — the banked voice becomes the voice of the person's AAC device: the tablet or speech-generating device they use to communicate.
Definition: AAC (augmentative and alternative communication) covers the methods and devices that supplement or replace natural speech — from simple communication boards to sophisticated eye-tracking tablets with synthesized voices. A banked personal voice makes AAC output sound like the person, not a generic robot.
Why it matters so much: voice is identity. Partners, children, and friends recognize a person by their voice — its warmth, humor, cadence. A generic synthetic voice communicates words; a banked voice communicates the person. Families consistently describe hearing the banked voice as profoundly comforting — "it still sounds like him."
When should you start voice banking?
As early as possible after diagnosis. This is the single most important advice, and the most commonly regretted delay. Voice banking requires clear, consistent speech — and ALS progression is unpredictable. People who feel "my speech is fine, I'll do it later" sometimes lose intelligibility faster than expected. Banking early costs nothing but a few hours; banking late may be impossible.
Practical guidance:
- Don't wait for speech symptoms. If diagnosis is confirmed, bank now — even if speech is completely unaffected.
- Bank before major procedures. Surgeries, hospitalizations, and illness can temporarily or permanently affect voice.
- Short sessions beat marathons. ALS causes fatigue — record in 15–20 minute sessions across days rather than one exhausting sitting.
- Re-bank if early. If you banked very early and speech later changes subtly, some programs allow supplementary recordings.
| Approach | Recording needed | Cost | Best for |
|---|---|---|---|
| Research/community projects (e.g., ModelTalker) | ~1 hour of scripted phrases | Free | Budget-conscious users comfortable with older tech |
| Commercial personal-voice services | 30–60 minutes of clear audio | Paid, varies | Most natural results with modern AI |
| Device-maker integrated banking | Varies by device | Often bundled | Users already in an AAC device ecosystem |
| AI voice-cloning platforms | As little as minutes of audio | Subscription | Tech-savvy users wanting latest quality |
How does the recording process work?
Preparation. Choose a quiet room — closets full of clothes work surprisingly well (soft surfaces absorb echo). Use a decent USB microphone or headset (not the laptop's built-in mic if avoidable). Have water nearby. Plan short sessions.
The scripts. Programs provide specific phrase lists designed to cover the language's sounds — read them as naturally as possible, at your normal pace and pitch. Don't "perform"; the goal is your everyday voice. Some modern AI approaches accept natural recordings (reading a book, telling stories) instead of scripted phrases — more pleasant, equally effective.
Session management. 15–20 minutes per session, with breaks. Stop when fatigued — tired voice recordings are inconsistent, and consistency matters more than volume. Most people complete banking across 3–6 sessions over one to two weeks.
Quality checks. Listen back to samples: consistent volume, no background noise, no clipping, natural pace. Re-record any session with lawnmowers, doorbells, or sirens in the background — the model learns whatever you feed it.
Storage and access. Store the finished voice model/files securely with backups, and make sure your care team and family know where they are and how to load them onto AAC devices. A banked voice nobody can find helps nobody.
How do families support voice banking?
- Make it a priority early. Gently — but firmly — encourage banking soon after diagnosis. This is the family's most time-sensitive contribution.
- Help with the technical setup. Microphone, software, quiet room, file management — remove every friction point.
- Record conversationally too. Beyond scripted phrases, record the person telling stories, laughing, talking to grandkids. Some modern systems can use conversational audio, and families treasure these recordings regardless.
- Capture signature phrases. The way they say "I love you," their laugh, their pet phrases — ask the banking program whether these can be incorporated; keep them as keepsakes either way.
- Respect the emotions. Voice banking forces confrontation with what's coming. Some people need encouragement; some need space; some need humor. Follow their lead.
- Plan for device integration. Work with the speech-language pathologist on which AAC device/software the banked voice will drive, and test the integration before it's urgently needed.
| Support role | What helps most | When |
|---|---|---|
| Spouse/partner | Encouragement, session companionship | Throughout |
| Adult children | Technical setup, file management | Setup phase |
| Speech-language pathologist | Program selection, device integration | Planning + integration |
| Friends/community | Normalizing the process, respite | Throughout |
| ALS Association chapters | Program guidance, equipment loans | Planning |
What if the person doesn't want to do voice banking?
Respect their autonomy — but make sure the reluctance is informed, not avoidant. Some people decline because the process feels like giving up; reframing helps: voice banking is preserving options, not conceding defeat. Others are overwhelmed by the diagnosis itself — in that case, offer to handle all logistics and revisit in a few weeks. What you shouldn't do is let treatable barriers (technical confusion, scheduling hassle, not knowing where to start) masquerade as a decision. Remove the barriers first; then respect whatever they decide. And document the decision either way — care teams need to know the plan.
What resources can help with voice banking?
- Your ALS clinic's speech-language pathologist — the single best starting point; they've guided dozens of families through this.
- ALS Association chapters — equipment loans, program guidance, and connections to others who've done it.
- ModelTalker (modeltalker.org) — the established free voice-banking program with years of ALS-specific experience.
- Commercial voice-cloning providers — for the most natural results; compare recording requirements and AAC compatibility.
- Online ALS communities — forums and groups where families share practical tips and emotional support.
Start with the speech-language pathologist — they'll match the approach to your specific situation, timeline, technical comfort level, family needs, and long-term device plans.
Frequently Asked Questions
When should someone with ALS start voice banking?
Immediately after diagnosis — don't wait for speech symptoms. Banking requires clear speech and ALS progression is unpredictable; delayers sometimes lose the window. It takes only a few hours across short sessions and costs little or nothing. Treat it as urgent, not eventual.
How much recording is needed for voice banking?
Older programs like ModelTalker need about an hour of scripted phrases; modern AI services typically need 30–60 minutes of clear audio. Quality beats quantity — 30 clean minutes outperform two noisy hours. Record in short 15–20 minute sessions to avoid fatigue.
What's the difference between voice banking and voice cloning?
Voice banking is the purpose (preserving your voice for future AAC use); voice cloning is the technology (AI building the replica). Older banking programs used different synthesis producing robotic results. When evaluating options, ask what the final voice sounds like and how much recording it needs.
Is voice banking free?
Several options are free: programs like ModelTalker cost nothing, and ALS Association chapters often help with equipment and guidance. Commercial AI services charge varying rates; some bundle with AAC devices. Cost should never be the barrier — start with free options and your care team.
Can you bank your voice after speech is already affected?
Mild speech changes may still allow banking — discuss with a speech-language pathologist. Significantly impaired speech usually can't produce a good model, since the AI learns the impairment too. This is why early banking matters. If banking is no longer possible, shift to message banking and selecting the best available synthetic voice.
What is message banking, and how does it differ?
Message banking records specific meaningful utterances — "I love you," laughter, pet phrases, personal stories — as fixed audio clips, rather than building a flexible synthetic voice. It complements voice banking: the synthetic voice handles everyday communication, while message-banked clips preserve the emotional moments exactly as spoken. Many families do both, and message-banked clips become treasured keepsakes regardless.
How does the banked voice work with communication devices?
The voice model loads into AAC software — tablets, dedicated speech devices, or eye-tracking systems — as the selected voice, speaking whatever is typed or selected. Work with a speech-language pathologist on integration before it's urgently needed, verify format compatibility with your AAC system, and keep backups of the voice files with the care team.
Related Guides
- AI Voice Credits Explained (2026)
- Free AI Voice Cloning: How It Works (2026)
- Urdu Voice Cloning Guide (2026)
Preserve the voice your family loves. Try SynVoice free and learn how voice preservation works today.