Write it, hear it, save it as an MP3. The voice is made here on your device by a small model that downloads once — nothing you write is ever sent anywhere.
Add to the list builds a reading a part at a time: write a bit, add it, drop in a silence, write the next bit, and read the lot when it is ready. Click a part to edit it. Anything still in the box when you press read goes on the end. With the list empty the box on its own is the reading.
Speed and pitch are asked of the voice itself, so changing either needs the reading made again. Everything under them re-lays the reading you already have, so they move while you listen.
The writing is split at sentence ends and each sentence is kept, so after an edit only what changed is voiced again. Headings, bullets and numbered steps are read as sentences of their own, and web addresses are skipped. Ctrl+Enter reads it aloud; space plays and pauses once a reading is ready.
the voice downloads once, then works with the network off