Convert any text to natural speech using Microsoft Neural voices — and download a perfectly synced SRT subtitle file with it. 20+ languages, free, no sign-up.
✨ New — Audio + SRT subtitles in one click
0characters
Language
Voice
Speed1.0x
Volume1.0x
Your browser does not support text-to-speech. Please try Chrome, Edge, or Safari.
How to use
1
Type or paste any text into the box above.
2
Choose a voice and adjust speed or volume if needed.
3
Click Speak to listen. Use Pause to pause mid-way, or Stop to cancel.
4
Tick "Also generate SRT subtitles" before downloading to get a perfectly synced .srt file alongside your audio.
Subtitles you can trust
✓
Millisecond-accurate timing. Timestamps are read directly from the voice engine as it speaks — not guessed afterwards by transcription.
✓
Word-for-word your script. The subtitle text is exactly what you typed. No speech recognition, so no misheard words or made-up sentences.
✓
Ready to upload. Standard .srt file that works with YouTube, Premiere Pro, CapCut, DaVinci Resolve and every major editor.
Frequently Asked Questions
Is this free?
Yes — completely free, no sign-up, no limits.
What voices are available?
All voices are Microsoft Neural voices delivered via Microsoft Edge TTS, with female and male options across 20+ languages. Quality is consistent on all devices.
Is there a character limit?
No hard limit. For very long texts, some browsers may split playback internally — just let it play through.
Do I need an internet connection?
Yes — an internet connection is required. Speech is synthesized using Microsoft Neural voices on our server. Your text is never stored or logged.
Is my text private?
Yes. Your text is sent to our server only to generate audio, then immediately discarded. We do not store, log, or share any content you submit.
How accurate are the SRT subtitles?
Exact. Unlike tools that transcribe the audio afterwards, the timestamps come from the voice engine itself while it generates speech, and the subtitle text is your original script word for word — no recognition errors, millisecond-level sync.
