An open-source Mac recorder transcribes calls and, using the user's uploaded files, suggests answers when it hears questions.
An open-source Mac recorder transcribes calls and, using the user's uploaded files, suggests answers when it hears questions.
The user adds PDF, text or markdown files to Parrot's knowledge base. These are indexed on the Mac. Parrot records system audio and the mic with no bot joining, and transcribes with Whisper locally. While the call runs, an assistant (Claude, Ollama or another model the user picks) shows answer cards that cite the source document. The user can pause it.
The site quotes about $0.07 per hour of calls with cloud cards and $0.00 fully local. It gives no accuracy measurements. Cards need a model, and cloud models receive transcript text.
It needs macOS 14 or later on Apple Silicon. Fully local mode uses Ollama, while cloud mode needs your own API keys, and transcript text goes to the provider you choose.
Build me a local tool that transcribes a live meeting from [audio source] and, whenever it hears a question, searches the documents in [folder path] and suggests a short answer with the source file named. Done means I can run it on a test recording and see the questions and suggested answers listed with timestamps. Keep everything on my machine and tell me before installing anything.
Score every demo request against your ICP, update HubSpot and ping the right rep in Slack within a minute of the form submit.
Every new Zendesk ticket gets a category, priority and sentiment tag plus a draft reply in an internal note before an agent opens it.
Copilot and Claude review each PR for bugs, missing tests and risky changes before a human looks, so reviewers can focus on design.