carlok — zsh — 88×30

cat _posts/2026-11-03-red-ink-sprint-24-a-voice-from-the-browser-autoplay-and-punches-thrown-and-landed-on-the-result-cards.md

red-ink sprint 24: a voice from the browser, autoplay, and punches thrown and landed on the result cards

The robot voice of sprint 23 is now the last fallback. Each word takes the first source that works: a clip listed in assets/voice/voice.js, then the browser’s own speech engine — a male voice of the game’s language, pitch 0.7, macOS novelty voices like Bahh and Fred skipped — then the formant synthesiser (b93cdd8). Nothing is bundled because Apple’s licence keeps the macOS voices to personal, non-commercial use and the Piper voices found were non-commercial or unclear, while the repo’s audio is claimed CC0; a Voice checkbox in the title settings and the pause menu turns it off, a voice list offers a Hear button, FIGHT is spoken about 0.2 s before the banner, and the banners themselves are localised (COMBATTI!, A TERRA). Regression stands at 31/31 with a stubbed speech engine; on this Mac the real list picks Daniel for English and Eddy for Italian, and how that sounds on an actual device is not something a headless probe can measure.

Autoplay lets the balance bot fight the player’s side (56a06c2). The checkbox lives in the title settings and the pause menu, is saved as sublevel-auto and also answers to ?auto=1; a move or attack key, the touch stick or buttons, or the pad takes over within the same frame, the next bout starts with autoplay again, cards still wait for the player, and an AUTO badge under the stamina bar says so in English and Italian. tools/autoplay-probe.mjs runs it in real time with no input: first action at 50 ms, Reyes won, Dag and Pike lost, every fight ended in a knockout within 16–38 s, and a J press stopped it in two ticks. The gamepad Back button now opens the card menu like Esc.

The win, lose and final cards gained a Punches row — “9 of 14 landed (64%) · 38.2 s” — with the defence line in Italian too, the counters living in G.log and never touching the simulation or its rng (ebf49c8). BACKLOG.md collects what the sprint left out, from an attract mode to a real playtest of the voice, and the offline zip now carries the voice folder; the closing run confirms regression 34/34, the bout table unchanged at 90.5 / 67.5 / 44.0 / 36.0 / 30.0 / 22.0, an identical step probe and no console errors in the environment probe (4a99bec, a142e47, #15).