Product & accessibility

Voice features: read-aloud and dictation in Autistic Mirror

Autistic Mirror can read responses aloud and accept voice input. Both are optional and can be turned off at any time. This article explains the neurological background, the technical setup, how the voices sound, and what happens to the data.

Why a voice feature reduces neurological load

Typing is not neutral. It couples three systems: fine motor control, visual feedback on the screen, and language production. In autistic processing, these often do not run silently in the background. Each one draws on executive resources. People who are already sensory-loaded have less budget left for them.

Reading is similar. Long blocks of on-screen text demand saccade control, contrast processing, and attention direction at the same time. On a day with high sensory-filter load, with migraine vulnerability, with POTS-related eye fatigue, or after a long masking day, exactly those resources are scarce. Hearing a response read aloud separates the content from the visual processing cost.

Dictating instead of typing reduces the motor and executive load of formulating. The thought is spoken once, instead of being structured in parallel with the act of writing. For many autistic and AuDHD people, this is the difference between a question that gets asked and a question that stays in their head.

Who notices the difference

The voice feature is not a comfort add-on. It lowers the entry barrier for several groups who often stay silent in text-only tools.

• AuDHD processing: dictation bypasses the inertia block in front of an empty input field.
• Co-occurring dyslexia or dysgraphia: typing carries a disproportionate cost.
• Eye fatigue on POTS, migraine, fatigue or post-viral days: on-screen reading becomes painful.
• Phases with high masking load: when the language-translation system is already overloaded, an additional modality helps.
• Motor co-occurring conditions like Ehlers-Danlos: keyboard sessions are physically limited.

What the feature actually does

There are two paths, and which one applies depends on the plan. In all subscription plans, and once as a trial in the free plan, read-aloud and dictation run through Autistic Mirror's servers. Once the one-time trial allowance in the free plan is used up, the browser's own voice features take over.

On the server path, the response text is sent to a speech model and comes back as an audio stream that starts playing while it is still being generated. For dictation, the recording is turned into text after you finish speaking, and the text appears in the input field. The recognition model detects the language by itself, so dictation also works in languages and browsers where built-in speech recognition is missing.

On the browser path, voice input uses the browser's Web Speech API and read-aloud uses its Speech Synthesis API. Which languages are recognised and how the voice sounds is then up to the device.

Both features are optional. Before the first dictation, a dialog explains which path currently applies and what is transmitted. Read-aloud can be started and stopped per response. There is no auto-play. Anyone who does not want to use the feature sees no difference from the classic text input.

How the voices sound

On the server path, the voice sounds natural and even, the same on every device. On the browser path, the sound depends on the system: flat on some devices, mechanical on others, usable on others again.

The server path costs money every time it is used. That is why read-aloud and dictation have separate allowances: once as a trial in the free plan, per billing month in the subscription plans. The settings show this month's usage. When an allowance in a subscription is used up, the app says so visibly instead of silently switching to the browser voice.

The voice feature was first available through the browser only. A present, plainly synthetic voice is usable. A missing voice is not usable. The server voice was added as soon as it was affordable, without any change to the feature scope.

Privacy

On the server path, the response text to be read aloud, or the audio recording of your dictation, goes through Autistic Mirror's servers to the Lovable AI Gateway and from there to the respective model provider. Nothing is stored: no recording, no audio, neither in the browser nor in the database nor in logs. All that remains is the recognised text in the input field. The data is not used for training.

On the browser path, Autistic Mirror transmits nothing. Whether the browser handles the processing on the device or sends text and audio to its vendor is up to the browser. Details on both paths, the providers and the legal bases are in the privacy notice.

Once confirmed, the recognised text is treated like any normal chat message and is subject to the same security and deletion rules as any other input.

A bright spot

Accessibility rarely arrives in one big launch. It arrives when features are released as soon as they are usable, and get better afterwards. The voice feature is a concrete example: rough at first through the browser, today with a natural voice through the server, honestly labelled, switchable at any time, with no pressure to use it. Anyone who needs it has it. Anyone who does not need it sees nothing change.

Autistic Mirror explains autistic neurology individually, related to your situation. For yourself, as a parent, or as a professional.