Skip to content

Memnon: A Local-First Voice Pipeline for Structured Knowledge #42

Description

@eitanfire

I wanted to figure out how I could dependably leverage my voice notes in AI-powered workflows. So I built a local pipeline that turns iPhone recordings into structured Obsidian notes, transcribed on your own machine with no audio sent to the cloud.

This talk walks through the architecture and live demo of Memnon: a Python pipeline small enough to read in an afternoon that goes from iPhone mic to AI-summarized Markdown note in under 60 seconds.

The stack: iOS Shortcut, iCloud Drive sync, whisper.cpp for local transcription, OpenAI or Ollama for structuring, and an Obsidian inbox note automated by a macOS launchd agent. A TypeScript MCP server layer lets Claude query your entire note history, a natural extension point for JavaScript developers.

Designed to be forked. Attendees leave with a working repo, a one-command setup script, and a clear surface for building their own AI knowledge workflows.

Repo: https://github.com/eitanfire/memnon

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    Approved ✅approved & confirmedTalk: Feature 🎦Propose a longer talk. Maximum 25 minutes.

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions