CVE-2026-108857: Hugging Face Text Embeddings Inference through 1.9.4 Cleartext API Key Logging
Hugging Face Text Embeddings Inference through 1.9.4 contains a cleartext logging vulnerability that exposes the configured apikey because the router's Args struct lacks a redact attribute for it. Attackers with access to router logs, container output, or OTLP telemetry can recover the Bearer token and call the protected embedding and rerank endpoints.
Affected Software
Event History
Frequently Asked Questions
Who can exploit this issue in practice?
An attacker needs access to router logs, container output, or OTLP telemetry that contains the router's startup logging. They can recover the configured Bearer token from those records and use it to call protected embedding and rerank endpoints.
What configurations are affected?
Deployments of Hugging Face Text Embeddings Inference through 1.9.4 with an api_key configured are affected. The exposure occurs because the router Args struct does not mark the api_key field for redaction.
How can I check whether credentials may already have been exposed?
Review router logs, container output, and OTLP telemetry for startup records containing the configured api_key or a Bearer token. Any accessible retained logs or telemetry containing that token should be treated as credential exposure.