Transcribe audio to text online

Transcribe audio to text online from interviews, meetings, lectures, podcasts, voice memos, and recorded calls. Hushscript is an online audio transcriber that can convert audio to text and returns editable, timecoded text you can search, correct, and export. Speaker detection separates voices where the source language supports it, and the editor lets you replace generic labels with real names. For a longer walkthrough, see the end-to-end audio transcription guide.

How to transcribe audio to text

To convert an audio file to text, add a local audio file, paste a supported audio link, or record up to 5 minutes in the browser. The preview always uses automatic language detection and shows up to 5 minutes before account creation; source-language, speaker, vocabulary, prompt, multichannel, and other recognition controls belong to the full app. Supported local audio can upload unchanged, while conversion runs in the browser only when required. Supported audio links are fetched and prepared server-side. Full recordings can run up to 10 hours; for multi-hour files, here’s how to transcribe a long recording.

In the app, choose automatic detection or set the source language, turn speaker detection off for a solo voice memo, or use multichannel handling when separate speakers already occupy separate channels. Add one-off keyterms and a custom prompt for names or jargon the sound alone cannot explain. Medical mode is optional and adds 100% of the recording duration to the billed minutes for that file.

Private audio transcription

Hushscript prepares local audio in your browser when required and sends only prepared audio for transcription. Stored transcripts, titles, speaker names, and translations are encrypted at rest. You decide how long a transcript stays: keep it, auto-delete after 7, 30, 90, or 365 days, or use private mode, which saves no transcript to your account. Private mode still uses online speech recognition.

Teach Hushscript your vocabulary

Keyterms and a custom prompt work once, for a single file. Saved dictionaries and prompt presets work every time you transcribe. Build a dictionary of the names, products, and jargon specific to your work, import it in bulk from a CSV or TSV file, and organize entries into categories with spelling variants attached to each canonical term. Save a prompt preset the same way, so the context you’d otherwise retype for every file is there by default. Mark a dictionary or preset as your default and it’s preselected automatically the next time you transcribe, so a standing setup doesn’t need reselecting every time. Both attach to a job alongside any one-off keyterms or prompt you add for that file alone, so a standing glossary and a same-day addition work together. They shape what the speech engine hears. Insights run later against the resulting transcript, so better vocabulary improves the source text but does not configure the Insights job.

Fix it fast

The editor is built for quick correction, not a full re-transcription. Undo any change, find and replace a misheard word across the whole transcript at once, rename a speaker so “Speaker 1” becomes a name, tag sections for later, or revert to the original if an edit goes wrong. When you choose Generate Insights, one optional free run returns everything at once: summaries, actions, decisions, open questions, quotes, chapters, speaker topics, subject clusters, export suggestions, terminology and profanity findings, and a rewritten, cleaned-up transcript. Compare it to the original transcript, edit it, or revert anytime. Regenerating later costs minutes, shown before you confirm.

Export anywhere

Download your transcript as TXT, SRT, VTT, CSV, TSV, Markdown, HTML, RTF, XML, JSON, DOCX, PDF, XLSX, ODS, ODT, ASS, TTML, SBV, LRC, FCPXML, or EDL. That’s 21 formats total, all unwatermarked. Not sure which one your next step needs? This guide walks through picking a transcript format. Export packs can bundle related document, subtitle, research, office, or editor formats in one ZIP. A password-protected .husharchive is a separate portable bundle with a manifest and checksums, and you can inspect one in the browser with the archive opener.

Batch audio transcription

One file at a time isn’t always the job, so batch transcription is built in. Stage several files at once and they sit in a queue until you hit start. Nothing transcribes the moment you drop it, so you can review and adjust options per file first, or start them all together. Related recordings (a split interview, a multi-part call, backup files from a recorder) can be merged into a single ordered transcript instead of staying as separate files you’d have to stitch together yourself. Your balance determines cost; active-job and service safeguards control how much work runs at once. If you’re clearing a whole directory of recordings, here’s how to transcribe a folder of recordings in one pass.

What Hushscript costs

Hushscript is pay-as-you-go, not a subscription: buy minutes in a pack and spend them on the projects you have. Packs run $1.99 for 45 minutes, $5.99 for 5 h (300 min), $12.99 for 15 h (900 min), $19.99 for 30 h (1,800 min), and $49.99 for 100 h (6,000 min). The largest pack works out to about $0.50 an hour. Minutes do not expire.

The free, browser-based tools convert and edit audio locally with no upload or account, but they do not transcribe speech. Audio to text transcription uses the preview and pay-as-you-go workflow described above; it is not an unlimited free converter.

Updated: October 26, 2026