Voice Capture Reconciler
Transforms iPhone voice notes into structured, reconciled updates within a Git-backed Markdown knowledge vault.
The Voice Capture Reconciler (VCR) skill automates the processing of Just Press Record iPhone voice captures directly into a personal Markdown vault without running persistent background servers or watchers. Using deterministic local Python helpers and hosted OpenAI speech-to-text models, the skill transcribes audio, isolates untrusted text, performs semantic reconciliation to map ideas to canonical owners, and strictly validates Git commits against configured remote branches before moving verified audio to the Mac Trash.
Key Features
01OpenAI-powered transcription utilizing secure macOS Keychain credentials
02Semantic reconciliation mapping voice captures into canonical vault documents and idea maps
03Deterministic audio discovery with SHA-256 deduplication and date-based folder cleanup
04Cryptographic remote Git branch verification before moving processed audio to Trash
052 GitHub stars
06Accidental recording detection tailored for iPhone Action Button misfires
Use Cases
01Reconciling Just Press Record voice captures into Obsidian or flat Markdown vaults
02Archiving auditable speech transcripts with strict Git provenance and automated cleanup
03Converting stream-of-consciousness voice memos into structured, non-committal idea logs