Hearlark Pro by CyberMaxMeeting notes from your recordings, made on your own laptop.
Drop a meeting, lecture or interview recording. Hearlark Pro writes the transcript, a summary, the decisions, the action items with owners and due dates and the open questions. Whisper runs inside your browser, so the recording is never uploaded.
- No upload
- No minute quota
- Up to 2 hours per file
- Word, Markdown, SRT exports
- Free up to 30 minutes
Notes work best in English; transcripts work in 90+ languages. Accuracy depends on the audio: check names and numbers against the recording.
Recent (kept in this browser only)
Pricing
Transcription runs on your device, so there are no per-minute fees and no minute caps.
Pro monthly
- Recordings up to 2 hours
- Full notes: summary, decisions, action items with owners and due dates, questions, topics
- Speaker names
- Word, Markdown and JSON exports
- Several files in one go
- AI summary polish
Secure checkout by Stripe. Your licence key appears right after payment; cancel any time from the billing portal. 14-day refund on request.
How it compares (prices read on each site, 1 Oct 2026)
| Hearlark Pro | Otter Pro | Notta Pro | |
|---|---|---|---|
| Monthly | $10 | $16.99 | $13.61 |
| Yearly | $84 | $8.33/mo billed yearly | $97.99 |
| Minutes per month | no cap | 1,200 | 1,800 |
| Where the audio goes | stays on your device | their cloud | their cloud |
| Joins live calls for you | no (bring a recording) | yes | yes |
Otter and Notta join your calls and work on any device without waiting; Hearlark Pro is for recordings you already have and keeps them private. Speed depends on your computer.
Questions
Is my recording uploaded?
No. The open Whisper speech model downloads to your browser once (from Hugging Face) and the recording is transcribed on your device. Notes are made on your device too. Only the optional "Polish the summary with AI" button sends the extracted notes (not the audio or the transcript) to our server, which answers and keeps nothing.
How long does a one-hour meeting take?
It depends on your computer. In our tests on a 4-core cloud CPU (no GPU), Whisper Tiny transcribed 60 minutes of audio in about 23 minutes, and Base did an 88-second meeting in 63 seconds including the model download. With WebGPU (recent Chrome or Edge) it is usually much faster. Keep the tab open; the notes appear when it finishes.
Does it tell who said what?
The transcript is split into turns at pauses. With Pro you name the speakers and tap a turn to set who spoke; the names flow into the notes, the action-item owners and the exports. Automatic voice recognition is not included.
How are action items found?
From what people say: "I'll send…", "Maria will draft… by Tuesday", "can you check… tomorrow?". Each item shows the owner and due phrase when the words name them, with a time stamp so you can check it.
Can I cancel?
Yes, any time from the billing portal link on your key page. Ask within 14 days for a refund.