GLM-TTS Test Console

checking…

Zhipu AI GLM-TTS — Zero-shot bilingual voice cloning (Chinese & English)

Text to Synthesize

Text (max 500 characters) — Chinese, English, or mixed

45 / 500

Voice Reference (Required)

🎤

Click to upload or drag & drop a WAV / MP3 file

GLM-TTS clones the speaker identity from a 3–10 second reference clip. Clear speech with minimal background noise gives the best results. Both Chinese and English references are supported.

Reference preview:

Reference transcript (optional — improves speaker similarity)

Leave blank if unknown. Providing the correct transcript significantly improves voice cloning quality.

Emotion

Intensity 50

Audio Adjustments

Volume (1–100)

Speed (-5 to 5)

Pitch (-5 to 5)

Authentication

API Key (if configured on server)