Loading vulnerability…
Unauthenticated credentialId abuse in Flowise text-to-speech generation endpoint
CVE-2026-41279 affects Flowise prior to version 3.1.0. The vulnerability is in the whitelisted, unauthenticated text-to-speech generation endpoint at POST /api/v1/text-to-speech/generate. This endpoint accepts a credentialId directly in the request body. When the endpoint is invoked without a chatflowId, Flowise uses the supplied credentialId to locate and decrypt a stored credential, such as an OpenAI or ElevenLabs API key, and then uses that credential to perform text-to-speech generation. The flaw is an authorization failure / user-controlled key reference issue in which an unauthenticated caller can cause the application to operate on a credential selected by identifier rather than enforcing proper ownership or access controls.
What it means and what to do now: patch path, interim mitigations, and the assume-compromise checklist.
What an attacker gets, and what they’ve been doing with it.
A remote unauthenticated attacker can invoke the text-to-speech generation functionality using stored third-party service credentials referenced by credentialId. This allows unauthorized consumption of victim-configured API credentials and service quota, potentially causing financial cost, resource exhaustion, and disruption of service availability for the affected organization or linked provider accounts. Based on the provided advisory data, the primary documented impact is availability-related rather than direct disclosure of the underlying secret value.
If you can’t patch tonight, do this now.
If immediate upgrade is not possible, restrict network access to the Flowise instance, especially the POST /api/v1/text-to-speech/generate endpoint, using reverse proxy rules, firewall policy, VPN, or IP allowlisting. Disable or block unauthenticated access to whitelisted endpoints where feasible. Monitor for unexpected text-to-speech requests and abnormal usage of linked provider accounts such as OpenAI or ElevenLabs, and rotate affected third-party API keys if unauthorized use is suspected.
Patch, then assume compromise.
Upgrade Flowise to version 3.1.0 or later, which fixes this issue. Ensure the text-to-speech generation endpoint enforces authentication and authorization checks before allowing use of stored credentials, and do not permit direct unauthenticated selection of credentials by credentialId.
No public exploits tracked yet. Mallory keeps watching.
Products and vendors Mallory has correlated with this vulnerability. Open in Mallory to drill down to specific CPE configurations and version ranges.
3 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.