San Francisco-based transcription company Tigerfish resolved a security lapse after exposing customer transcriptions, raw audio and video files, and internal invoices on an Amazon-hosted cloud server accessible to anyone without authentication.
The exposed data included transcriptions of police interviews and body-cam footage from law enforcement agencies and Silicon Valley tech companies. Invoices detailing client relationships and service usage were also compromised.
Security researcher Anurag Sen discovered the files in October on an Amazon S3 bucket listed with a predictable web address and no access controls. He alerted Tigerfish to the vulnerability but received no response. After sharing details with a media outlet, the company made direct contact.
Tigerfish secured the S3 bucket following inquiries, blocking public access to all files and directory listings. Adam Goldberg, Tigerfish president, confirmed the company verified all other S3 buckets were similarly secured. Goldberg attributed the initial delay to Sen's email being caught in a spam filter.
Tigerfish did not respond to questions about whether affected customers have been notified of the exposure, raising questions about incident response protocols and customer transparency.
Sen, who operates from Indonesia and has spent a decade identifying exposed datasets online, has previously discovered sensitive U.S. military emails on Defense Department servers, customer record databases, national license plate surveillance systems, and SMS code streams used for account logins.
For Tigerfish, the exposure presents material business risk. Law enforcement and tech companies that handle classified or proprietary information demand stringent data protection from vendors. Client churn among these high-value accounts—where switching costs are low if trust erodes—could directly threaten revenue. In a market where security is a core competitive moat, a publicized misconfiguration undermines differentiation.


