Speech to text
Speech to text for recorded audio
Hushscript turns an existing audio or video recording into an editable transcript with speaker labels. It sends only prepared audio for recognition, removes its transcription audio after processing, and never trains on your recordings. Preview up to 5 minutes before signing in, then transcribe the full file with prepaid minutes. It works on recorded media, not live dictation.
Drop an audio or video file to preview your transcript
First 5 minutes preview
Choose a fileVideo stays local. Transcription audio isn't kept.
Paste a shared link to an audio file. How link import works
Record up to 5 minutes with your microphone. Nothing uploads until you choose to transcribe.
No sound detected. Check your microphone.
Saved in your browser's recording format. Need MP3 or WAV? Convert it with Hushscript's free audio converter
Check your email for your sign-in link. It's valid for 15 minutes.
This preview stays on this page and isn't carried into the app. Sign in and choose your file again; a new personal account's first upload gets its own free preview of up to 5 minutes. Full transcription needs enough available minutes. Eligible personal accounts can receive 30 free minutes once after successful card verification, or with their first purchase if not already granted.
How it works
- 01
Add a recording
Choose an audio or video file, paste a shareable audio link, or record up to 5 minutes in the preview widget. Files can run up to 10 hours.
- 02
Check the preview
Review up to 5 minutes of speaker-labeled text before you create an account. Only the preview clip is sent at this stage.
- 03
Transcribe and edit
Sign in, upload the full recording, and correct the transcript in the editor. Export it in any of 21 formats when it is ready.
At a glance
- Price
- Packs from $1.99, about $0.50 per hour of audio on the largest pack
- Transcription length
- Up to 10 h per file
- Upload limit
- No file size limit
- Languages
- 99 languages
- Privacy
- Hushscript removes its transcription audio after processing, never trains on your recordings, and encrypts stored transcripts at rest.
What is speech to text?
Speech to text converts spoken words in a recording into written text. Hushscript works with files you already have: add a recording, check a real preview, then edit the transcript instead of replaying the same passage to type it by hand. Hushscript is built for recorded conversations and monologues, not a keyboard microphone that types live.
Speech to text software for recorded files
To convert speech to text, add an audio or video file, paste a shareable audio link, or record up to 5 minutes in the preview widget. A transcript makes the recording searchable and gives you text tied to speakers and time. Correct a name, find a phrase, rename a speaker, or return to the original text if an edit goes wrong.
The same workflow fits an interview, meeting, lecture, field note, voice memo, or recorded presentation. Files can run up to 10 hours. Developers building their own workflow can use the speech-to-text API, while the page here is for working with recordings in the Hushscript app.
Private speech to text for recordings
Hushscript is cloud speech to text built around privacy. For a local video, it prepares the audio in your browser, so the source video stays on your device. Only prepared audio goes for recognition, and stored transcripts are encrypted at rest. You decide how long a transcript stays, or use private mode to save none to your account.
AI speech to text that keeps conversations readable
AI speech to text turns a multi-person recording into labeled speaker turns instead of one block of text. You can rename a label such as “Speaker 1” after you recognize the voice, then export a transcript that still shows who said what. Speaker detection is available across the supported transcription languages. Machine transcripts can contain errors, so check names and speaker labels against the recording.
For a solo recording, you can turn speaker detection off. For audio with separate channels, multichannel handling can keep those channels distinct. Saved dictionaries, prompts, and one-off keyterms can also give the speech engine context for names, acronyms, and recurring terminology.
Free speech to text online, then pay as you go
The free speech to text preview shows up to 5 minutes of your own recording before signup. Full transcription needs an account and enough available minutes. Past the free minutes, prepaid packs start at $1.99 for 45 minutes and there is no subscription.
You spend minutes on completed transcription. The first Insights generation for a transcript uses no minutes and can add summaries, actions, decisions, chapters, and a cleaned-up transcript; a later regeneration costs minutes and shows the charge before confirmation.
Speech to text facts
| Question | Hushscript |
|---|---|
| Best fit | Existing audio or video recordings, not live dictation |
| Preview | Up to 5 minutes before signup |
| Recording length | Up to 10 hours |
| Languages | Roughly 99, with automatic detection |
| Speaker labels | Available across supported transcription languages |
| Editing | Correct text, rename speakers, find and replace, undo, redo, and revert |
| Exports | 21 formats with no watermark |
| Billing | Pay-as-you-go, no subscription |
See all supported source languages on the languages page. Need the broader file workflow? Use the audio to text converter. If your speech is still happening, record it first with the free browser voice recorder, save a copy, then bring that file back for transcription.
Updated:
Why Hushscript
Private by design
Only prepared audio is sent, transcription audio is deleted after processing, recordings never train models, and stored transcripts are encrypted at rest.
Speakers kept clear
Speaker detection separates who said what in supported languages instead of returning one block of text.
No subscription
Eligible accounts get 30 free minutes once, after card verification or with a first purchase. Then buy prepaid minutes only when you need them.
Frequently asked questions
How does speech to text work?
Speech to text AI, also called automatic speech recognition (ASR), analyzes the words in recorded audio and returns timestamped text. Hushscript prepares your file, transcribes it with speaker labels where supported, then gives you an editor and 21 export formats.
Can I use speech to text free online?
Yes, to start: preview up to 5 minutes before signing in, and eligible accounts get free starter minutes. After that it is pay-as-you-go.
Can I get video speech to text?
Yes. For speech to text from video, add a local video file and Hushscript prepares its audio in your browser, so the source video stays on your device. Only the prepared audio uploads for transcription.
Does Hushscript do real time speech to text?
No. Hushscript transcribes recordings you already have, not live speech. To capture live speech, record it first with the free browser voice recorder, then bring the file here for transcription.
Can I use Hushscript as a speech to text translator?
Yes, for recorded audio or video. Choose one or more supported target languages before transcription, and Hushscript produces translated transcript text in the same job; each target adds 25% of the recording duration to billed minutes. It does not create translated voice or translate live speech.
What does cloud speech to text mean?
Cloud speech to text sends audio to an online speech-recognition service instead of running the recognition model entirely on your device. For local files, Hushscript prepares media in your browser and uploads only prepared audio; shareable audio links are fetched and prepared server-side. The transcription-audio copy is removed after the transcript is saved, and recordings are not used to train models.
Which languages does the AI speech to text support?
Hushscript auto-detects and transcribes roughly 99 spoken languages. Speaker detection is available across the supported transcription languages.
Can I convert sound to text?
Yes. Hushscript works as a sound to text converter by turning spoken dialogue in recorded audio or video into an editable transcript. It does not transcribe music, background noise, or sound effects, and it does not generate a voice from text.
Which transcript formats can I download?
You can export 21 formats, including TXT, DOCX, PDF, Markdown, SRT, VTT, CSV, JSON, and editor timeline formats. Exports have no watermark.
Start with 30 free minutes
Preview up to 5 minutes before you create an account, and a new personal account gets a free preview of its first upload in the app. Full transcription needs an account and minutes; eligible personal accounts get 30 free minutes once, after card verification or with a first purchase. No subscription.
Get 30 free minutes