0.15

Error code 3221225477

This error is more likely to occur on PCs with 6 GB of RAM.❗
After restarting, launch VoicegerN (this workaround is only effective once; a restart is likely required each time you generate a model).
The error is caused by insufficient RAM. Restarting frees up runtime memory and resets the page file state.
Additionally, on the first run, antivirus software or Windows Defender may scan model and DLL files, which can consume a large amount of memory.

OSError: Page file too small or MemoryError

Why this error occurs 🤔

If there is insufficient memory (for example, if your PC has less than approximately 12.5 GB of RAM), the pagefile.sys file will grow in size.
If you have 6 GB of RAM and your C: drive has more than approximately 15 GB of free space, there's no need to worry. However, if your C: drive has less than approximately 10 GB of free space, you may encounter this error.

The system's page file is usually located at C:\pagefile.sys. When RAM runs low, the system uses this file as cache (if you're unsure, you can ask an AI)😊
You can search online for guides on how to change the location of the Windows page file.

If your C: drive is a mechanical hard drive (HDD), frequent use of the page file as cache may shorten the drive's lifespan. If you are using an SSD, you can ignore this note.🙏

Note

Tested on Windows 11. Under default settings, the page file will automatically shrink when you are not generating voice output.

Caution ❗

If you wish to reduce the page file size or change its location, please look up relevant guides yourself. The author accepts no responsibility for any issues that may arise from such operations.

0.17

When Using VoxCPM2 TTS with Mixed‑Language Input (Chinese, Japanese, Cantonese), Zundamon Reads It as Japanese

There is currently no solution for this issue. This TTS model does not have a language selection option, so it tends to mix up Chinese, Cantonese, and Japanese, reading them as the wrong language.
For mixed‑language text, you will need to split it into separate segments and generate each part individually; otherwise, the output may be incorrect.

NOTE

Even when a language option is available, it often misreads Chinese as Japanese and Japanese as Cantonese... (this is a bit frustrating, to be honest)

The Program Shows a "missing runtime file" Error

Did you manually switch the TTS core in the settings?
If so, switch it back: Settings > Software > TTS Core > Select Qwen3TTS > Confirm.
If that is not the case, please download the full TTS core files and replace them.
If the issue persists, please report it via the Contact section. 🙏