Global Tongue Twister Voice Warm-up Speed Challenge Read aloud, get real speech-recognition accuracy, and race your personal best

Step 1: Choose a language and tongue twister

Pick a language set and a real local tongue twister. Mature SpeechRecognition languages enable true speech scoring.

華語(繁體台灣)cmn-Hant-TW

吃葡萄不吐葡萄皮,不吃葡萄倒吐葡萄皮。

先慢慢分清「葡萄」與「皮」,再追求速度。

Step 2: Record aloud or use timer mode

Microphone permission

After you start, your browser may ask for microphone access. Speech recognition runs through the native Web Speech API; this page calculates text similarity and time locally.

Your browser does not support SpeechRecognition, so only timer mode is available and no speech score is shown. Try the latest Chrome or Edge.
0.00
Press start when ready

Practice tips and privacy

Start slowly and clearly, then increase speed. Tongue twisters reward clarity and control, not speed alone.

If the page shows timer mode, the browser or language cannot provide reliable speech recognition, so no fake score is displayed.

Personal bests are stored only in localStorage on this device and disappear when browser data is cleared.

What is the Global Tongue Twister Voice Warm-up Challenge?

This tool turns real tongue twisters from around the world into a browser game: choose a prompt, read it aloud, and supported browsers use the native Web Speech API to capture a transcript and score accuracy against the target text. Learn the background of tongue twisters, the browser Web Speech API, and the SpeechRecognition interface.

True scoring, not just a stopwatch

It is useful for presenters, teachers, podcasters, streamers, actors and language learners who want a quick vocal warm-up or party challenge. In supported languages, the tool compares the recognized transcript with the target phrase using edit-distance similarity, then combines accuracy and elapsed time into a score.

Honest fallback when speech recognition is unavailable

If the selected language or browser is not reliable for SpeechRecognition, the page clearly switches to timer-only mode and says there is no speech score. You can still warm up, self-rate, and save the fastest local time.

Privacy and related tools

Other useful tools: