Voice banking is the process of recording your natural speaking voice so it can be used to build a synthetic version of it later. If disease, surgery, or injury eventually takes away your ability to speak, a banked voice lets you continue communicating in something that sounds like you, rather than a generic computer voice. The technology is most commonly used by people with progressive neurological conditions, but it’s increasingly accessible to anyone with a smartphone.
Who Needs Voice Banking
The people most likely to benefit are those facing conditions that gradually destroy speech. Motor neurone disease (MND), also called ALS, is the most well-known example. As the disease progresses, the muscles controlling speech weaken until spoken communication becomes impossible. Other progressive neurological conditions, including Parkinson’s disease and multiple sclerosis, can also affect speech over time, though the timeline and severity vary widely.
People scheduled for surgery on the larynx, tongue, or throat (often due to head and neck cancers) also bank their voices beforehand. In these cases, the loss of natural speech can be sudden rather than gradual, making early planning essential. Some parents of young children with conditions that may affect future speech also explore voice banking, though the technology works best with a mature, stable voice.
How the Recording Process Works
Voice banking involves reading a set of scripted phrases into a microphone. The software analyzes the recordings to capture the unique acoustic characteristics of your voice: pitch, tone, rhythm, and the way you shape vowels and consonants. It then uses those characteristics to generate a text-to-speech voice that can say anything, not just the phrases you recorded.
The number of phrases you need to record depends on the service. Acapela requires a minimum of just 50 phrases to create a synthetic voice. ModelTalker recommends at least 400 phrases, suggests 800 for better quality, and allows up to 1,600. More recordings generally produce a more natural-sounding result, because the software has more raw material to work with. A session of 50 phrases might take under an hour, while recording 800 or more could spread across several days.
For the best results, you want a quiet room with minimal echo. A USB condenser microphone plugged into a laptop is the most practical setup for home recording. Professional studios use higher-end equipment, but most voice banking services are designed to work well with consumer-grade gear. Consistency matters more than perfection: recording in the same room, at the same distance from the microphone, at a similar energy level across sessions.
Voice Banking vs. Message Banking
These two terms often come up together but serve different purposes. Voice banking creates a flexible synthetic voice that can speak any new sentence typed into a device. Message banking records specific phrases, sentences, or expressions exactly as you say them. Think of message banking as saving audio clips: “I love you,” a favorite joke, the way you call the dog.
Many people do both. The synthetic voice handles everyday communication, while the message-banked clips preserve the emotional texture of phrases that matter most. Some services allow “double dipping,” where the same recordings count toward both your synthetic voice build and your message library.
Major Services and Their Costs
Several dedicated platforms offer voice banking, each with different pricing and features.
- Acapela (My Own Voice): Roughly €99 per year as a subscription, or around €999 for an outright purchase. People with MND in the UK can access it free through MND Association funding.
- SpeakUnique: Offers a voice build for £175, a voice repair option (for voices already showing deterioration) at £250, and a voice design service for £300. Also free to people with MND through the same funding pathway.
- ModelTalker: A longstanding service developed through research at a US university, historically offered at low or no cost for people with ALS.
- ElevenLabs: A newer AI-driven platform with an Impact Program that provides free voice cloning to people with qualifying conditions. Team Gleason, a nonprofit supporting people with ALS, helps connect applicants to this program.
Costs vary by country, and nonprofit organizations can often cover them. In the US, Team Gleason is one of the most prominent sources of funding for voice preservation technology and related equipment. The Department of Veterans Affairs also provides support for eligible veterans.
Apple Personal Voice and Built-In Options
Starting in 2023, Apple built voice banking directly into its operating systems, making it available to millions of people at no extra cost. The feature, called Personal Voice, lets you create a synthetic version of your voice by reading a series of prompts on your device. It works on iPhone 12 or later running iOS 17, iPad Air (5th generation) or later with iPadOS 17, and Macs running macOS Sonoma 15 or later.
The process takes about 15 minutes of reading and then several hours of on-device processing (usually overnight while the device charges). The resulting voice integrates with Apple’s Live Speech feature, which lets you type text and have it spoken aloud in your own voice during phone calls, FaceTime, or in-person conversations. For someone who wants a basic banked voice without navigating specialized services, this is the most straightforward option available today.
How Banked Voices Connect to Communication Devices
A synthetic voice is only useful if it works with the tools you’ll actually rely on. People who lose the ability to speak typically use augmentative and alternative communication (AAC) devices. These range from dedicated hardware with eye-tracking or switch controls to tablet apps that generate speech from typed or selected text.
Most voice banking services export your finished voice as a file that can be loaded into compatible AAC software. Compatibility varies by service and device, so it’s worth checking before you start recording. Speech-language pathologists who specialize in AAC can help match a voice banking service to the device setup that best fits your needs, both now and as your condition progresses.
When to Start
The single most important piece of advice from clinicians and people who have been through the process: start as early as possible. Voice banking works best when your speech is still clear and strong. Once a condition begins affecting speech, even subtle changes in articulation, breath support, or vocal quality will carry over into the synthetic voice.
For people diagnosed with ALS or MND, this means banking your voice soon after diagnosis, ideally before any noticeable speech changes. The disease can progress unpredictably, and waiting too long is one of the most common regrets reported by patients and families. A first-person account published in a 2025 medical journal urged neurologists and general practitioners to raise voice banking at the point of diagnosis rather than waiting for speech symptoms to appear.
Some services, like SpeakUnique’s voice repair option, can work with speech that has already started to deteriorate, using AI to reconstruct a clearer version. But the results are better when the starting material is stronger. If you or someone you know has received a diagnosis that could affect speech, the time to look into voice banking is now, not when speaking becomes difficult.