Hearlark

Private speech-to-text. Whisper runs inside your browser: your audio is never uploaded. Free, no sign-up.

Drop an audio or video file here, or click to choose MP3, WAV, M4A, OGG, WEBM, MP4… anything your browser can play.

Model Language Task Engine

Choose a file to start. The model downloads once and is then cached by your browser.

Why it's private: the page downloads the open Whisper model from Hugging Face once, then does all the listening on your own device. The audio file is decoded in your browser and never sent to any server, including ours. Close the tab and it's gone.

Tips: Tiny handles clear speech fast; pick Small for accents, noisy rooms or non-English audio. WebGPU (recent Chrome/Edge) is usually much faster than CPU; long files on CPU take a while, so keep the tab open. Accuracy varies with audio quality, so check names and numbers before publishing.

Recorded a meeting? Hearlark Pro turns it into notes on your own device: summary, decisions, action items with owners and due dates, speaker names and a Word file. Free up to 30 minutes; Pro $10/month.

More free tools from CyberMax
All CyberMax tools Article text extractor demo MCP & API kit for AI agents

Need transcripts, text or data from hundreds of web pages instead? CyberMax's pay-per-result web tools run on the Apify Store, callable from Make, Zapier, n8n, Python or any AI agent.