Shared speech for the Parris app familyA voice, worth keeping.
Record once. Reuse across adventures, learning and gentle movement. Complete phrases first, carefully assembled cues when needed.
198 words + 12 reusable phrases 550 starter clips planned Search & review catalogue →
Plan before generating
The starter catalogue is a plan, not 550 finished recordings.
250Common words
120Numbers & directions
80Feedback clips
100Full phrases
6,083 script characters before any voice-specific credit modifiers. Start with ten clips and listen before generating the rest.
Open the generation notebook
Checks the loaded Sylvie voice pack. No speech is generated.
Compare the playback modes
Same-voice Naomy cues: try a complete recording, then an intentionally segmented count. Device speech uses a different voice.
Word joining is useful for counts; listen for the breaks between words. No neural or paid request is made here.
New · Grammar-based sentence studioRandom sentence generator
Mix verbs, subjects and objects into real sentences. Every generated word must have a saved recording in the selected voice pack. Grammar rules prevent random word salad; audio is only available where recordings exist.
Loading starter catalogue…
Compare original words, smoothed words and recorded phrases
Use the same sentence to compare. Phrase recordings usually sound most natural.
12 reusable phrases have been recorded in Naomy’s voice. View the next phrase priorities.
Green words have a standalone recording in the chosen voice. Amber words are only in the planned catalogue. Complete saved recordings always take priority. The phone fallback speaks the whole sentence in your device voice, not Naomy.
Optional neural speech uses a separately configured local engine. This demo has no neural model, paid API key or automatic speech fallback.