Little Words: Jessica spoken picture names

Provider: ElevenLabs Text to Speech. Voice: Jessica. Model: Eleven v4.
Generated on 15 October 2026 under a paid Starter plan.
50 distinct word recordings serve the exact displayed labels of 57 cards.
The elephant recording is reused from the owner-selected third audition voice.
The original MP3 files are served unchanged; no runtime generation is used.
Source metadata and hashes are retained in source-assets/audio-jessica/.

Incorporated into Little Words under the ElevenLabs Creative Platform Terms:
https://elevenlabs.io/creative-terms

16 October 2026 expansion:
100 additional distinct picture-name recordings, using the same owner-selected
Jessica voice, Eleven v4 model, English, stability 0.5 and similarity 0.75.
The combined game has 150 distinct spoken labels across 157 cards.
All new recordings are pre-generated. One interrupted motorcycle download was
recovered from the existing ElevenLabs history result without another generation;
its compact delivery copy is encoded locally. Other new speech originals are
served unchanged. Originals, hashes and technical checks are retained separately
in source-assets/audio-expansion-100/. Individual-word listening review is pending.

16 October 2026 hog label correction:
The wild-boar picture now uses the exact displayed word "hog", recorded with
Jessica in Eleven v4 through the ElevenLabs web interface. One Generate action
created two automatic variants; the selected first variant cost 3 credits and
the unused second variant cost 0. No further generation was used in processing.
The selected MP3 is retained unchanged: decoded peak 0.539653, below the 0.75
speech ceiling. The game now has 151 distinct spoken labels across 157 cards.
Original, generation identifiers, hashes and signal checks are recorded in
source-assets/audio-hog-correction/. Listening review remains pending.