ai-elements
Warn
Audited by Socket on Oct 13, 2026
1 alert found:
SecuritySecurityscripts/speech-input.tsx
MEDIUMSecurityMEDIUM
scripts/speech-input.tsx
The code implements ordinary browser-based speech transcription and shows no evidence of malware or intentional malicious behavior. The primary security issue is exposing an OpenAI API key through NEXT_PUBLIC_OPENAI_API_KEY in client-side code, which can lead to credential theft, unauthorized API usage, and unexpected charges. A server-side proxy with protected credentials and request controls should be used.
Confidence: 99%Severity: 72%
Audit Metadata