FAQ
Frequently asked questions about Prismical, the open-source AI note taker.
General
What is Prismical?
Prismical is a free, open-source AI note taker. It captures system audio in the background to transcribe meetings, lectures, and voice notes — without a bot joining your call. AI generates structured notes, summaries, and action items automatically.
How is this different from Otter, Fireflies, or Fathom?
Those tools send a bot into your meeting that everyone can see. Prismical captures system audio in the background — participants never know it's there unless you tell them. Your audio also stays on your device instead of being sent to external servers.
How is this different from Granola?
Granola is also bot-free, but it's closed source, costs $14/month, and sends your transcripts to cloud AI for processing. Prismical is open source, free, and can process everything locally with on-device models.
What meeting platforms does Prismical work with?
All of them. Prismical captures system audio, so it works with Zoom, Google Meet, Microsoft Teams, Slack Huddles, WebEx, Discord — any app that plays audio through your computer.
Is Prismical free?
Yes. Prismical is open source under the MIT license. Download it, use it — no cost, no credit card, no usage limits.
Privacy & Technical
Does any audio leave my device?
Not by default. Local models (Whisper, Parakeet) process audio entirely on your device. If you choose to use cloud providers (Deepgram, OpenAI, Assembly AI), audio is sent to those providers using your own API keys.
What AI models does Prismical use?
For local transcription: Whisper and Parakeet, both running on-device. You can also connect cloud providers like Deepgram, OpenAI, and Assembly AI with your own API keys.
Can I use Prismical for things other than meetings?
Absolutely. Use it for voice memos, lecture notes, interview transcription, brainstorming, or quick note-taking. AI structures whatever you capture into clean, organized notes.
What is the MCP server?
The MCP (Model Context Protocol) server lets you connect Prismical to AI assistants like Claude and ChatGPT. They can read your meetings and notes, search past conversations, and write notes on your behalf.
What languages are supported?
Prismical supports transcription in English, Spanish, French, German, Portuguese, Japanese, Chinese, and more — depending on the model. Whisper supports 90+ languages.
Do I need a powerful computer?
Local transcription works best on machines with modern hardware (Apple Silicon Macs, or Windows PCs with a decent GPU). On older hardware, you can use cloud providers instead.
Can I export my data?
Yes. All notes, transcripts, and summaries can be exported as JSON or Markdown.
Contributing
Can I contribute?
Yes. Prismical is MIT licensed and actively maintained on GitHub. We welcome contributions — features, bug fixes, documentation, and translations. Join our Discord community to get started.
Is there a mobile app?
iOS and Android apps are in private beta. Join the waitlist to get early access. Mac and Windows desktop apps are available now.