The vulnerability resides in the _conversation function of the server/backend.py file in ramon-victor freegpt-webui. A manipulation of the argument model removes authentication checks, allowing an attacker to invoke the backend conversation API without credentials. This weakness is categorized as CWE‑287 (Authentication Bypass) and CWE‑306 (Missing Authentication Mechanism). The result is the potential for an attacker to send arbitrary conversation requests, potentially revealing private data or gaining control over user sessions.

The affected software is the ramon‑victor freegpt-webui project, which follows a rolling release model. No specific version numbers are available for the vulnerable or patched releases because the project maintains continuous delivery. The flaw only exists in installations that are no longer supported by the maintainer, so currently unsupported or legacy deployments are at risk.

The CVSS score is 6.9, indicating a moderate severity of the unauthenticated remote risk. The EPSS score is not available, but an exploit has been publicly disclosed, meaning that exploitation is feasible with the right knowledge. The vulnerability is not listed in the CISA KEV catalog. Attackers can remotely manipulate the model argument to bypass authentication and use the conversation API, subject to the discovered exploitation conditions. The likelihood of exploitation is uncertain but potentially high for unattended or legacy deployments.