Voice Notes to Structured Personal Notes via Telegram & Local LLaMA

This n8n workflow powers a Personal Notes Assistant that captures voice memos or text notes sent via Telegram bot. It processes inputs offline on devices like Raspberry Pi 5, using local Whisper for accurate audio transcription into structured text.

The transcribed or direct text notes are then analyzed by LLaMA 3.2 via Ollama, which generates summaries, suggests tags (e.g., mood, theme), and adds metadata like timestamps. Notes are securely stored in a local database such as SQLite or JSON, enabling semantic search, topic-based retrieval, and historical queries like 'Summarize stress notes from last week' or 'Show travel-tagged entries.'

Benefits include complete privacy with no cloud dependency, efficiency for busy users, and reflective AI prompts to deepen journaling. Use cases span personal productivity, idea capture during commutes, professional brainstorming, or therapeutic reflection—all automated and accessible via simple Telegram chats.