VoxNotch garden-planning demonstration Captured October 28, 2026 This is an original, harmless demonstration, not a customer recording or case study. INPUT garden-planning.wav: 16.4 seconds of synthetic English speech, generated with eSpeak NG 1.52.0 (en-us voice, 145 words per minute). script.txt: the original spoken script. ACTUAL PRODUCT OUTPUTS garden-planning.md: downloaded through the VoxNotch export dialog, then opened outside the application. Its transcript text was not manually corrected. garden-planning.json: the product's JSON transcript export, including word timing. garden-planning.srt: the product's subtitle export. garden-planning-notes.md and garden-planning-notes.json: the product's note exports. The note was written manually and linked to source time 12.44 seconds. Clicking the note's displayed 0:12 marker was checked against the audio playhead. CONFIGURATION Real faster-whisper 1.2.1, Whisper small, CPU, int8; voice activity detection on. The model was downloaded before the final run; inference used the local cache. The demo used temporary local files, an in-memory store and an inline job queue. It does not demonstrate persistent storage or a separate worker deployment. The synthetic input voice is not a VoxNotch voice-generation sample. WALKTHROUGH /assets/product/garden-planning/walkthrough.mp4 The silent walkthrough uses actual browser footage, trimmed between steps, with still holds on the transcript and opened export. Elapsed time is not a benchmark. Captions: /assets/product/garden-planning/walkthrough.vtt