Get an AI meeting summary without uploading your meeting data. Choose local transcription and a local AI model. No bot joins your call.
Source-available · local-first · backed by Y Combinator
Transcribe enabled microphone and system audio on your computer with local Whisper.
Keep permitted slides, documents, and app history alongside the conversation.
Choose a local AI model to draft notes without sending the transcript to a hosted model.
Screenpipe captures enabled microphone, system audio, and screen sources. Use local Whisper for transcription and Ollama for private meeting notes, summaries, and action items. It works with Zoom, Google Meet, Teams, and other apps with audio. Cloud providers and sharing are optional, separate data paths.
screenpipe runs on your device. choose the plan that fits your workflow.
01
great for getting started
$25
/ month7-day free trial, then $25/mo
Existing lifetime licenses keep working.
02
great for individuals or small teams
$50
/ seat / monthfor power users and small teams
A private note taker needs to keep the whole meeting workflow local: recording, transcription, summary generation, and storage. With Screenpipe, you choose the transcription engine and AI provider separately.
Local models use your computer's memory and processing power. Review the transcript, names, and action items before using the notes.
Set up fully local meeting notesRecording and consent requirements still apply when no bot joins the call.
Granola, Grain, and Fireflies also offer botless capture. The privacy question is where your audio, transcript, and summary are processed and stored. Their cloud workflows may suit teams that prioritize shared notes and CRM integrations.
Read the data-flow guide before enabling cloud AI, team sync, or a connected app.
Yes. Choose local transcription, run a summary model through Ollama, and save the notes on your computer. Keep cloud AI, cloud transcription, sync, remote media analysis, enclave processing, and external exports disabled for that workflow. If you select a hosted summary model, it receives the transcript even when the recording is local.
Screenpipe records the microphone and system audio you enable. A transcription engine turns the audio into text, then your selected AI model drafts the summary and action items. Choose local Whisper and a local model for on-device processing, or explicitly choose a hosted provider. Review names and commitments against the transcript.
Yes. Screenpipe captures enabled system audio and microphone sources without joining as a meeting bot. Test permissions and both audio sources before relying on it for a call, and follow your meeting's recording and consent requirements.
After downloading the transcription and AI models, local transcription and summaries can run without internet access. Test a short recording and summary offline on your machine. Model quality, tool support, speed, and memory needs vary; hosted AI and a remote meeting itself still require a network connection.
Screenpipe supports local transcription, local summary models, and searchable screen context across apps. Granola, Grain, and Fireflies offer cloud meeting workflows, and all have botless capture options. Compare where audio is transcribed, where summaries are generated, and where notes are stored. A botless recorder or private cloud storage alone does not mean on-device AI.
Only if Screenpipe was already recording the required audio with the right permissions and devices enabled. It can summarize retained recordings, but cannot recover a conversation that was never captured.
Source-available · local-first · backed by Y Combinator
download screenpipe freecompare with other tools →get updates on screenpipe news and releases
