Skip to main content
screenpipe automatically transcribes all audio from your meetings, calls, and conversations. everything runs locally using Whisper for maximum privacy.

​
how it works

  1. audio capture: screenpipe records from your microphone and system audio
  2. speech-to-text: audio is transcribed using Whisper (runs locally)
  3. speaker identification: different speakers are automatically labeled
  4. searchable storage: transcripts are indexed and searchable

​
setup

​
basic setup

# start screenpipe with audio enabled (default)
screenpipe

​
capture both mic and system audio

# list available audio devices
screenpipe --list-audio-devices

# capture specific devices
screenpipe --audio-device "MacBook Pro Microphone" --audio-device "BlackHole 2ch"

​
choose transcription engine

# local (default) - most private
screenpipe --audio-transcription-engine whisper-large-v3-turbo

# cloud - faster, requires API key
screenpipe --audio-transcription-engine deepgram

​
query transcriptions

import { pipe } from "@screenpipe/js";

// find discussions about a topic
const results = await pipe.queryScreenpipe({
  q: "budget review",
  contentType: "audio",
  limit: 10
});

// get transcripts from today's meetings
const today = new Date().toISOString().split('T')[0];
const meetings = await pipe.queryScreenpipe({
  startTime: today,
  contentType: "audio"
});

​
using the meeting pipe

for automatic meeting summaries:
  1. install the meeting pipe from the store
  2. it automatically detects when meetings start
  3. generates summaries with action items
  4. can sync to Notion, Obsidian, or other tools

​
transcription quality tips

  • use a good microphone
  • reduce background noise
  • whisper-large-v3-turbo gives best accuracy
  • add --language en if you only speak English (faster)

​
speaker identification

screenpipe automatically identifies different speakers: 
const results = await pipe.queryScreenpipe({
  contentType: "audio"
});

for (const item of results.data) {
  if (item.type === "Audio") {
    console.log(`Speaker: ${item.content.speaker_id}`);
    console.log(`Text: ${item.content.transcription}`);
  }
}

​
privacy

  • all transcription runs locally on your device
  • audio files stored in ~/.screenpipe/data/
  • no audio is sent to the cloud (unless you choose deepgram)
  • you can disable audio recording with --disable-audio

​
next steps