Skip to content
ttsmaker

Speech to text

Upload an audio or video file, or press the microphone and talk: ttsmaker turns speech into text for free, without an account. Whisper large-v3-turbo recognizes about 99 languages, and you can download the result as TXT, SRT or VTT.

Drop an audio or video file here

MP3, WAV, M4A, OGG, WEBM, MP4, MOV

Transcript

Your transcript will appear here

How to convert speech to text

  1. 1.Choose File or Dictation

    Use the File tab for a recording you already have. Use the Dictation tab to speak into your microphone and see the words appear.

  2. 2.Set the language

    For files, Auto lets Whisper find the language by itself. For dictation, pick the language you are going to speak.

  3. 3.Add the file or start talking

    Drop the file into the box or press the microphone. A file is recognized piece by piece, and the text grows as each piece is done.

  4. 4.Copy or download

    Copy the text, or save it as TXT. Choose SRT or VTT if you need subtitles with start and end times.

Key facts

  • Cloud model: OpenAI Whisper large-v3-turbo, hosted on Cloudflare Workers AI.
  • About 99 languages, with automatic language detection for files.
  • 30 minutes of free cloud transcription per IP address per day; the counter resets at 00:00 UTC.
  • Past those 30 minutes, Whisper base keeps working inside your browser with no daily cap.
  • Export formats: TXT, SRT and VTT, with timestamps in the subtitle files.

Which speech-to-text tool do you need?

Every page below uses the same engine, but each one is set up for a different job. Pick the row that matches what you have and what you want to get.

You have You want Go to
A recording: interview, lecture, voice memo, podcast Text with timestamps Audio to text
A video file from a phone, camera or screen recorder The words spoken in it Video to text
Your own voice and a microphone Text while you speak Voice typing
A video you want to caption An SRT or VTT file Subtitle generator
A file and an interest in the model itself OpenAI Whisper without installing it Whisper online

File or microphone: what changes?

Files go through Whisper. Your browser decodes the recording, turns it into 16 kHz mono sound and cuts it at quiet moments into pieces no longer than 30 seconds. Each piece is recognized, and the pieces are joined back with their timestamps, which is why a file can become subtitles.

Dictation uses the speech recognition built into Chrome, Edge and Safari, so words appear while you are still talking. In Chrome this recognition is done by Google’s service: that is how the browser works, not our server. Firefox has no built-in recognition, so the tool records short pieces between your pauses and sends them to Whisper instead.

What this tool does not do

  • No speaker labels. Everyone in the recording ends up in one stream of text.
  • No links. Paste a YouTube or podcast address and nothing happens: download the file first.
  • No meeting bot. Nothing joins Zoom or Meet for you; record the call and upload the recording.
  • No translation yet. Text comes out in the language that was spoken.
  • No API. The tool works in the browser only.

The daily cloud minutes and the speed cap are explained on the limits page, and the full table of supported audio and video formats shows what each browser can open. For a language-by-language view, see all languages.

Frequently asked questions

Is this speech to text really free?

Yes. There is no account, no watermark and no paid plan (November 2026). The cloud gives each IP address 30 minutes a day, and after that the browser model carries on for free.

How accurate is the transcription?

It depends mostly on the recording. Clear speech close to the microphone comes out well; loud music, people talking over each other and heavy noise cause mistakes. We do not publish an accuracy percentage because it changes from file to file.

Is my audio saved anywhere?

No. We do not keep audio or transcripts on our servers. In cloud mode, short 16 kHz pieces pass through our server to Cloudflare Workers AI and the text comes back; in browser mode the audio never leaves your device.

Does it work on a phone?

Yes, in a mobile browser. Files are decoded in the phone's memory, so a very long recording can be slow or fail on a phone with little memory. Splitting the recording into parts usually solves it.

Which languages can it recognize?

Whisper covers about 99 languages and detects the language of a file automatically. Accuracy is higher for widely spoken languages and lower for languages with less training data.

Sources

Updated: