CVE-2025-49841 GHSL-2025-053: GPT-SoVITS Deserialization of Untrusted Data vulnerability
Attackers can execute arbitrary code via unsafe deserialization in the GPT-SoVITS WebUI when loading user-controlled model paths.
- Exploited in the wild
- Not known
- Login needed
- Not needed
- Attackable
- Over the internet
- Victim action needed
- Not needed
- Fix
- None listed
What is CVE-2025-49841?
CVE-2025-49841 is a high-severity unsafe deserialization vulnerability in RVC-Boss GPT-SoVITS, published on 15 Jul 2025. An attacker can exploit it over the internet without logging in, with no action from the victim.
Official description
GPT-SoVITS-WebUI is a voice conversion and text-to-speech webUI. In versions 20250228v3 and prior, there is an unsafe deserialization vulnerability in process_ckpt.py. The SoVITS_dropdown variable takes user input and passes it to the load_sovits_new function in process_ckpt.py. In load_sovits_new, the user input, here sovits_path is used to load a model with torch.load, leading to unsafe deserialization. At time of publication, no known patched versions are available.
Weakness: CWE-502
Which versions are affected by CVE-2025-49841?
It affects 1 product.
Products include Rvc-Boss Gpt-Sovits-Webui.
How do I fix CVE-2025-49841?
No fix has been published for CVE-2025-49841 yet. Limit exposure of the affected software until one is available.
Upgrade, and only deserialise data from trusted sources or with an allow-list of types. More about Unsafe deserialization.
Is CVE-2025-49841 being exploited?
There is no known exploitation of CVE-2025-49841 in the wild and no public exploit code.
Other vulnerabilities in RVC-Boss GPT-SoVITS
- CVE-2025-49833 GHSL-2025-045: GPT-SoVITS Command Injection vulnerability
- CVE-2025-49834 GHSL-2025-046: GPT-SoVITS Command Injection vulnerability
- CVE-2025-49835 GHSL-2025-047: GPT-SoVITS Command Injection vulnerability
- CVE-2025-49836 GHSL-2025-048: GPT-SoVITS Command Injection vulnerability
- CVE-2025-49837 GHSL-2025-049: GPT-SoVITS Deserialization of Untrusted Data vulnerability