voice-memo-to-obsidian-selfhosted

add condense transcription function, audio titling, don't have newlines for transcript, add transcript override for faster feedback cycles for ai prompt tweaking, save last ai response for easier debugging

add note about viewing vtt files in obsidian

replace gemini with speaches-ai for transcription and ollama for analysis

improve: more context in readme, and cleanup prompt vs format

Initial commit: voice memo to obsidian pipeline