Prismical
Open-source AI meeting note-taker that transcribes locally, without a bot joining your call.
🔗 Visit Prismical
Description
Prismical listens to your meetings, lectures, or voice memos and turns them into organized notes with action items — like having a colleague quietly taking notes for you — except it runs the transcription on your own computer instead of sending your audio to a company's servers, and no bot has to awkwardly join your Zoom call to do it.
It captures system audio directly (works with Zoom, Teams, Meet, Slack, WebEx or any audio source), transcribes with local Whisper models, and can also connect to Claude, ChatGPT or Gemini via MCP for summarization. It's open source (MIT license) and free, with both a fully local mode (no account needed) and an optional cloud mode for cross-device sync.
💬 Our review
The short version: Prismical is a strong pick if the "no bot in my meeting" and "nothing leaves my device" angles matter to you more than polish — it's free and genuinely local-first, which most competitors aren't.
Granola, Otter and Fireflies are more mature, better-designed, and have deeper calendar/CRM integrations, but they're paid subscriptions and typically process audio in the cloud (or, in Granola's case, still require the app to be actively listening in a specific way). Prismical's local Whisper transcription means privacy-sensitive users — legal, medical, journalism — can use it without a data-processing agreement, and being free removes the cost calculus entirely. The tradeoff is a much smaller project (74 GitHub stars) with less refined summarization quality and fewer integrations than the established paid players.
💰 Pricing
📊 Global score
🤖 AI-enriched data
Free forever, local mode requires no account
Pros
Free and open source
Local-first, no bot in the meeting
Works with any meeting platform
MCP integration with major AI assistants
Cons
Small project, less polished
Fewer integrations than paid competitors
Summarization quality less proven