Integrating Minimaxi Voice Synthesis

After the v3.81 patch, the built-in Minimaxi (the parent company of Conch AI) text-to-speech feature has been added, supporting dozens of characters and 15 languages. It also allows adjustments for emotions, pitch, and more, making it quite usable.

Registration and login URL: https://platform.minimaxi.com/login

After logging in, real-name verification is required using a bank card number and the mobile number registered with the bank.

Then, open the software and navigate to Menu → TTS Settings → Minimaxi API.

For users in China, select api.minimaxi.com in the API field. If you wish to use Minimaxi's international services, choose api.minimax.io. Please note that although both are Minimaxi services, one is the domestic site (https://www.minimaxi.com/audio) and the other is the international site (https://www.minimax.io/audio). Accounts, API keys, balances, etc., are not interchangeable between the two.

Fill in the interface key in the SK field. You can create one at: https://platform.minimaxi.com/user-center/basic-information/interface-key

Note: If you have not passed real-name verification, the test may fail.

Additionally, when using this method, please go to Menu → Tools/Options → Advanced Options → Voice Adjustment and set the "Simultaneous Voice Synthesis Count" to 1 and the "Pause After Voice Synthesis" to a value greater than 25. Otherwise, you may exceed the rate limit and encounter failures. Regular users are only allowed 3 requests per minute, which is one request every 20 seconds.

Emotion Selection for Pronunciation

Minimaxi supports 7 emotions: Happy, Sad, Angry, Fearful, Disgusted, Surprised, Neutral. However, testing shows that the differences are minimal. If needed, you can adjust these settings in the interface.

Finally, note that unless you have a corporate account with a higher tier on Minimaxi, you may frequently encounter rate limit errors. Set a longer pause time in the advanced options under "Pause After Voice Synthesis" to mitigate this.