No audio yet
Click the name to change it. The respelling and the IPA fill in by themselves — click either one to correct it.
Where
What
The IPA and respelling fill in by themselves as you type. The first one downloads eSpeak (~9 MB) and says so while it runs; after that it is instant.
Filling automatically from the name and language. Anything you type yourself stops being overwritten.
Paste this into your site
Where you will put the downloaded files. Any static folder works, and another domain works too — audio elements don't need CORS headers.
Assumes the component is already on the page:
<script type="module" src="https://cdn.jsdelivr.net/npm/@jailson/say-my-name@0"></script>
Don't want to record your own voice?
That's a reasonable position — a clean, isolated sample of your voice on a public page is exactly what voice-cloning models want, and you don't owe anyone one. A synthesized voice you picked deliberately and labelled honestly is a perfectly good answer.
What this page can do itself
One engine runs here: eSpeak NG, in this tab. Speak it fills in the IPA, generates the audio, and ticks Synthesized voice for you. The voice is robotic, but it is derived from real letter-to-sound rules rather than guessed at.
The voice dropdown beside each pronunciation picks between its 104 voices — same synthesizer, different vocal tract. Since the result is a file you download and host, the voice you pick is the voice every visitor hears, unlike browser speech synthesis where you get whatever that device has.
What you'd go elsewhere for
Nothing below runs here. These are tools you'd open yourself, generate a clip in, and bring back with File — the file is then read in this tab like any other, and still never uploaded anywhere.
| Tool | Voice | Uses your IPA? | Account? |
|---|---|---|---|
| phoneme-synthesis | Robotic (eSpeak) | Yes — paste IPA directly | No — runs in your browser |
| Kokoro · TTS.ai | Natural, open models | No — type a respelling until it sounds right | No |
| Azure Speech Studio · Google Cloud TTS demo | Natural | Yes — paste SSML <phoneme> |
Yes, free tier |
Only the eSpeak-based ones connect your IPA to the sound; the natural-voice tools ignore it, so you tune those by ear. Robotic-but-exact often beats smooth-but-wrong for a name.