Paste text or drop in a PDF, Word file, EPUB or article link, pick one of 322 neural voices across 75 languages, and get an MP3 back, plus SRT and VTT subtitles timed to each word. No account. No paid tier. 7,000 characters per voiceover, refilling every two minutes. Documents are parsed in your browser rather than uploaded, and transcription runs on-device, so the file stays with you and only the text you choose to speak gets synthesised.
I built this because every free text-to-speech site I tried wanted an account before it would let me hear anything, capped me at 500 characters, or gave me a robotic voice and no download.
ReadAnyText.com is the version I wanted: paste or import, 322 voices, MP3 and WAV out, and word-accurate SRT/VTT alongside the audio. That last part is the bit I'm most pleased with, the subtitles line up per word, not per paragraph.
Three things I'd rather say up front than have you discover:
7,000 characters per voiceover, refilling every two minutes. A novel is a lot of passes. Cancelled and failed renders cost you nothing.
Documents are parsed in the browser. The PDF or EPUB never leaves your machine. The text you actually press play on does go out for synthesis. I'd rather state that plainly than let "privacy-first" imply more than it should.
Happy to answer anything, and genuinely after the awkward feedback more than the nice kind.