Meetily Brings Free Meeting Transcription Without Upload or Subscription

Meeting assistants have emerged as a popular AI application, promising to record, transcribe, and summarize calls automatically. The appeal is straightforward: no more manual note-taking. However, existing services have two significant drawbacks. Most charge between $10 and $20 per month, and they typically require users to upload meeting recordings to cloud servers—a privacy concern for anyone handling confidential material.

Meetily is designed to sidestep both problems. Built on open-source AI models, it runs entirely on your computer rather than in the cloud. The free community version is available to download from GitHub, with no subscription required. A Pro version exists for $10 per month, but the core functionality is available at no cost. Setup is straightforward: install the application on Windows or macOS (Linux users can build it from source), and Meetily will offer to download and install the AI models that perform the transcription and summarization work.

Once those models are installed, you grant Meetily access to both your microphone and system audio—a crucial permission that allows it to capture all participants in a call. Because it records at this level rather than integrating with a specific platform, Meetily works with any video conferencing service: Zoom, Google Meet, Microsoft Teams, or any other. You can even use it for in-person meetings if your laptop's microphone can pick up the room. When you start a call, hit the record button; during the meeting, you can watch the transcript appear in near-real time or minimize the window and focus on the conversation. When the meeting ends, stop the recording to view the full transcript and request an AI-generated summary, optionally providing instructions on which points to emphasize.

A beta feature extends this capability to previously recorded files: you can drag an audio or video file into Meetily to get a retroactive transcript and summary, making it useful for transcribing phone recordings or in-person conversations captured on your phone.

The free version does have limitations. Most notably, it cannot identify which person said what in a transcript (no speaker labels), a feature that many professional users will find essential. The author could not locate a free open-source tool that consistently provides speaker identification. Other alternatives exist—noScribe is free and can identify speakers but only processes files, not live meetings; Anarlog offers a sleeker interface and access to multiple AI models but requires more setup effort; Quill provides free local transcription but mandates account creation, undermining the offline-first appeal. The landscape is fragmented, suggesting the market for free meeting transcription is still early and competitive options will likely proliferate.