Meetily, an open-source meeting transcription and summarization app, is offering a free alternative to the $10–$20 per month cloud services that dominate the meeting-assistant category. The tool runs locally on Windows and macOS, uses open-source AI models bundled at install time, and never sends call audio to a remote server. A $10 per month Pro version exists, but the community build on GitHub is fully functional and free.
The pitch is straightforward: transcription has quietly become one of the most useful applications of modern AI, and the underlying models are now good enough that a laptop can do the job without help from a datacenter. Services like Otter, Fireflies, and the built-in transcribers in Zoom and Microsoft Teams charge a monthly fee and require uploading meeting recordings — a non-starter for anyone handling confidential material like legal work, medical calls, or unreleased product discussions.
Meetily sidesteps that by capturing microphone and system audio directly from the operating system. That approach means it is not tied to any specific videoconferencing platform. It works with Zoom, Google Meet, Microsoft Teams, and any other tool that plays audio through the computer. In principle it also works for in-person meetings, provided the laptop microphone can pick up every voice in the room.
Key facts
- 01Meetily is free and open source on GitHub, with a $10/month Pro tier — versus $10–$20/month for typical cloud meeting assistants.
- 02The app runs on Windows and macOS; Linux users can build from source.
- 03It records mic and system audio directly, so it works with Zoom, Google Meet, Microsoft Teams, or any other call software.
- 04Transcription and summarization run locally using bundled open-source models — no cloud upload required.
- 05The free tier lacks speaker labels; competing local tools include noScribe, Anarlog (formerly Hyprnote), and Quill.
Installation follows a two-step flow. The app downloads the AI models needed for speech-to-text and summarization on first launch, then asks for permission to record both mic input and system audio. From there, a single record button starts capture. Users can watch the transcript build in near real time or minimize the window and focus on the call. When the meeting ends, the app produces a full transcript and, on request, an AI-generated summary with optional guidance on which topics to prioritize.
A beta feature extends the tool beyond live calls: users can drag in an existing audio or video file and receive the same transcript-and-summary output. That covers the common case of recording an in-person conversation on a phone and processing it later on a laptop.
The free version has one meaningful gap. It does not label speakers, so the transcript reads as an undifferentiated stream of text rather than a tagged back-and-forth. For interviews, depositions, or any multi-participant meeting where attribution matters, that is a real limitation. No fully free and open-source tool consistently solves speaker diarization at the local level today, which is why several adjacent projects have carved out narrower niches.
noScribe is a free tool that does identify speakers, but it only processes existing recordings — it cannot capture a live call. Anarlog, previously known as Hyprnote, offers a cleaner interface and support for a wider set of AI models, though setup involves more configuration. Quill provides free local transcription but requires account signup, which undercuts the appeal of an offline tool and pulls users into upsell flows.
The wider category — cloud meeting assistants at $10–$20 per month — has grown quickly on the back of remote and hybrid work. Pricing has held firm because the incumbents bundle transcription with summaries, action-item extraction, CRM integrations, and searchable archives. Meetily is not trying to match that feature set. It is doing the two things most users actually need — a transcript and a summary — and doing them without a subscription or a cloud round-trip.
The caveat that applies to every AI tool applies here too. Transcripts are not perfect, and auto-generated summaries reflect the model's judgment about what mattered in a meeting, which is not always the user's judgment. For high-stakes notes, the transcript is best treated as source material for a human-written summary rather than as the final artifact.
The economics of local AI tools are quietly getting more attractive. As open-source speech models improve and consumer laptops handle more inference on-device, the value of a $15/month cloud transcription subscription narrows to the collaboration and integration layer — not the transcription itself. Meetily is one of the first meeting-focused apps to demonstrate that plainly. Expect the cloud incumbents to respond by leaning harder on integrations and team features, because the raw transcript-and-summary loop is no longer something they can charge a premium for.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
