Comparison
Harknotes vs Otter.ai
Otter.ai sends a bot into your meetings and uploads your audio to the cloud. Harknotes transcribes locally — no bot, no cloud audio, no compliance risk.
The key difference: where your audio goes
Otter.ai works by joining your meeting as a participant. It records the call from inside the meeting room, uploads the audio to Otter’s servers, and transcribes it there.
Harknotes captures audio directly from your operating system. No bot joins your meeting. No one else can see a recording indicator. Your audio is transcribed on your machine using whisper.cpp — a local AI model.
Feature comparison
| Feature | Harknotes | Otter.ai |
|---|---|---|
| Bot joins your meeting | No | Yes |
| Audio sent to cloud | No | Yes |
| Works offline | Yes | No |
| Open source | Yes (AGPL-3.0) | No |
| Linux support | Yes | No |
| Price | Free & open source | From $16.99/mo |
| Speaker labels | Yes (local) | Yes (cloud) |
| Markdown export | Yes | No |
Privacy
Otter.ai’s privacy policy states that audio recordings are stored on their servers. This creates risk for:
- NDA-protected conversations — legal, financial, and HR meetings
- Internal security policies — many orgs prohibit sending meeting audio to third parties
- Client calls — many clients object to being recorded by a third-party service
Harknotes’ approach: audio never leaves your machine. Summaries run against a local model by default — or your own API key, sent directly to the provider you choose. There’s no middleman server at all.
Pricing comparison
Otter.ai’s paid plans start at $16.99/month. Harknotes is free and open source (AGPL-3.0) — no tiers, no usage limits, no account. You bring the AI: a local model via Ollama, or your own API key.
Who should use Harknotes instead of Otter.ai?
- Teams with compliance or security requirements
- Anyone who can’t have a recording bot visible in client meetings
- Linux users (Otter has no Linux app)
- People who want their notes as Markdown files
More comparisons