By ShinobiTools Team · Last updated: July 2026
Turn speech into text for free: an interview, a voice memo, a recorded call, a lecture.
Upload the recording and the words come back as text you can copy, search and edit — no account, no watermark, no per-minute fee. It works as a sound file to text converter and an audio file to text converter for any common format.
No account. 10 recordings a day. Files up to 60 minutes here, and it resets at midnight.
Transcribe audio now →Most "free" transcribers cap you at a few minutes a day or stamp a watermark on the output until you pay. ScribeGrab runs on our own GPU, so there's no cloud bill per minute — 10 recordings a day, individual files up to 60 minutes and 1.5 GB, about 100 languages auto-detected, and your upload is deleted the moment it's done.
You don't pick the language: it's detected from the recording itself, and the transcript comes back in the language that was spoken. Want it in another language, that's a separate button afterwards, with the timings kept intact — see translate subtitles.
The AI is a speech recognition model on our own GPU, not a rented API, which is why speech to text AI here has no per-minute fee. With this speech to text free online tool the transcript itself is always free; a paid pass only adds longer files and extras. Unlike Google's Voice typing in Docs, which needs a live microphone, it works from a file you already have.
Windows speech to text, Android speech to text or a Mac: it's the same browser tab, with no desktop app and no speech to text Chrome extension to install. On a phone, upload the voice memo or a WhatsApp voice note and get text back; old dictaphone recordings work the same way. If iPhone's own dictation is the problem (speech to text iPhone not working is usually a settings issue), record a voice memo as usual and upload the file here instead.
There's no direct Excel or Word export on the free tier: you get plain text that you paste into cells, a document or an email. A speech to text Google Docs habit works the same way, just paste the result in. And a full written transcript is for deaf and hard-of-hearing readers as much as for speed.
The honest way to pick a speech to text tool is to read its transcript of your own recording: accuracy depends more on the recording than the brand, which is why our accuracy breakdown focuses on recording conditions. MP3, a phone memo or a video file are all the same upload here, and to transcribe a recording for free you need no account. A Zoom recording works the same way as any other file; see transcribing a Zoom meeting for the speaker-label version if it was a call with several people. Coming from Restream's built-in transcription? Same idea, without the streaming platform wrapped around it. How to convert speech to text here takes four steps:
Alongside the plain text you get SRT and VTT, where every line carries a start and end time. That's what makes it possible to jump straight to a moment in a two-hour recording instead of scrolling. If several people speak, speaker labels mark who said what.
Formats: MP3, WAV, M4A, FLAC, OGG, AAC and the common video files. No converting first.
On clear speech it's close to word-perfect. Strong accents, people talking over each other or a noisy room are harder for every tool, so read those back once. We wrote an honest breakdown of how accurate AI transcription really is. Punctuation is inferred from the pauses and the grammar, and so a long unbroken stretch of talking still arrives split where a person would have breathed.
Yes. 10 recordings a day, no per-minute charge, no watermark and no account. Files up to 60 minutes and 1.5 GB each. That free speech to text tier usually beats competitors' free allowances, and the full TXT, SRT and VTT export is included.
About 100, detected automatically from the recording, including Cantonese, Czech, Bengali, Dutch, German and Spanish. You don't have to choose.
Yes. Plain text as TXT, and SRT/VTT with a start and end time per line.
One sentence per line, with punctuation added automatically, and "Speaker 1:", "Speaker 2:" labels when more than one voice is heard. It's one continuous transcript: splitting it into speech to text notes by topic is a manual step afterwards.
It's deleted the moment the transcript is done, and the results are wiped within 45 minutes (24 hours with any pass, free or paid).
Yes, that's exactly what it does: upload the audio from your phone's voice memos or a voice recorder and download the free transcript as TXT, SRT or VTT a few minutes later. It replaces typing up an interview by hand.
Speech to text turns a recording into written words. Text to speech does the opposite: it reads written text aloud for a listener rather than a reader. ScribeGrab handles both: audio to text and text to speech.
Close to word-perfect on clear audio, dropping off with heavy background noise, crosstalk, whispering or a thick accent. See the accuracy breakdown for how we measured it.
No. It works on recording files: record first with your phone or any app, then drop that file here for the text. If you're wondering how to use talk to text for live dictation, that's your phone's or Word's own dictation button; for a file you already have, this talk text page needs no keyboard at all once the recording exists. Voice typing in Google Docs is the live kind too.
No. Speech to text on Microsoft Word is Word's own built-in dictation button, and Copilot's dictation lives inside Microsoft's apps too; both need a live microphone. This page is separate: upload a recording and paste the transcript into your document, which is how speech to text Word users get a file transcribed.
No download and no speech to text Chrome extensions: speech to text for PC here just means opening this speech to text website in a browser, uploading the file and getting the transcript back. Speech to text Windows users get works the same way on Mac, Linux, a Chromebook or a phone.
ChatGPT and Claude are chat interfaces first, not built for long recording files. ElevenLabs charges past a small free allowance and needs an account; Evernote's transcription lives inside its note-taking app. This page is free up to 60 minutes per file, with no account and no note app.
No. It runs on our own GPU with our own speech recognition, not a rented cloud API from Microsoft or Amazon, which is also why the free tier can be generous. Buzz, by comparison, is an open-source desktop app you install yourself.
Yes. Generate the SRT file here, then import it as captions in CapCut's Text panel or on Resolve's subtitle track, like any standard subtitle file.
Only enough to upload the file; after that it processes on our server. Keep the tab open until the upload finishes, or you'll have to start it again.
No. It's free either way; file length and the number of recordings a day are the only limits. With voice speech that's quiet or whispered, getting closer to the mic helps more than anything else.
ShinobiTools · KvK 42147182 · info@shinobitools.com