Introduction
Prismical is a free, open-source AI note taker for meetings, voice notes, and local-first transcription.
Prismical is a desktop note-taking app for meetings, lectures, interviews, and personal notes. It captures microphone and system audio from the apps you already use, transcribes the audio with local or connected cloud models, and keeps the transcript attached to the note you are writing.
Use it when you want AI notes without inviting a meeting bot, routing every call through a hosted note taker, or moving your workflow into a browser tab. Notes, transcripts, folders, tags, favorites, and skills live in the same workspace, so capture and cleanup happen in one place.
Local-first, cloud-optional
Use downloaded Whisper models when you want transcription to run on your machine. Connect cloud transcription providers when that better fits your speed, accuracy, or deployment needs.
Quickstart
Install the desktop app, grant permissions, download a local model, and create your first note.
Meeting notes
Record a call, review the transcript, and turn the conversation into structured notes.
Personal notes
Capture voice notes, drafts, ideas, and private notes outside a meeting.
Workspace
Learn the sidebar, notes list, folders, tags, favorites, and settings areas.
How Prismical Works
Capture audio
Start recording from a note, the recording dock, or a meeting alert. Prismical captures microphone and system audio from your computer, so it works with Zoom, Google Meet, Microsoft Teams, Slack Huddles, WebEx, Discord, browser calls, and anything else that plays audio through your computer.
Transcribe with your chosen model
Use downloaded Whisper models for local transcription, or connect cloud transcription providers when you prefer their speed, accuracy, or operating model.
Review the transcript
The transcript stays connected to the note. Open the transcription panel to follow speaker-labeled segments, copy the transcript, or add transcript text into the note.
Use skills when needed
Skills can clean up, enhance, append to, or rewrite note content. Use local or connected language models for that text-generation step.
What To Read Next
Transcription
Choose models, record audio, review speaker-labeled transcript segments, and add transcript text to notes.
Recording dock
Start, stop, monitor audio levels, and open transcription controls from the note surface.
Skills
Use AI skills to clean up notes, enhance rough text, append sections, or rewrite the current note.
Integrations
Connect Prismical context to MCP-compatible assistants and local workflow tools.
Prismical is MIT licensed and open source at github.com/amicalhq/prismical. You can download the desktop app for Mac or Windows, or join the community on Discord.