Local desktop app that records, transcribes, and summarizes meeting audio using Whisper and local LLMs.
Capture desktop audio, transcribe with Whisper, and query meeting notes without sending data to the cloud.
Pensieve is a local-first desktop application designed to record, transcribe, and summarize audio from meetings and applications directly on your computer. It suits legal professionals who handle sensitive client calls and require strict confidentiality without sending audio to external vendors.
Pensieve runs in your system tray to capture application audio, generating local transcripts using Whisper. It builds searchable meeting summaries using a local Ollama instance or OpenAI, and includes a built-in chat view and MCP server to query your meeting archives.
It suits privacy-conscious legal practitioners, corporate counsel, and researchers needing offline meeting records without cloud compliance risks. It does not suit teams looking for automated cloud calendar bots or multi-party cloud recording workflows out of the box.
Converts captured meeting audio into text directly on your device using whisper-cpp and FFmpeg.
Summarizes meeting transcripts using a local Ollama LLM instance or optional ChatGPT API key.
Connects your local transcript repository directly to external tools using the Model Context Protocol.
Lets you search and ask specific questions across past meeting recordings directly within the app.
Pensieve is free to download and run locally. The only potential external cost occurs if you choose to connect an optional paid OpenAI API key for summarization instead of using local Ollama.
876 open remote AI-training roles, updated November 15, 2026; most pay $20–$170 an hour. Zekai lists them free; the application goes to the employer’s platform.
No, if you select local Whisper for transcription and Ollama for summarization, all audio processing and notes remain entirely on your local machine.
It features an interactive Chat view within the desktop app and includes a built-in MCP server to share context with external tools.
See Zekai first in Google