DetectifAI opens its voice deepfake detector to developers

DetectifAI opened its voice deepfake detector to developers on October 19th through an API and an MCP server, founder Tarini Sai Padmanabhuni announced, advertising a 24-millisecond check time. Padmanabhuni said the system caught 97.3% of cloned voices in an independent benchmark, while the same benchmark recorded 94.47% accuracy and an 8.4% false-positive rate on genuine clips. Developers weighing voice-security tools must balance the low latency against those false alarms.

DetectifAI opens its voice deepfake detector to developers Founder Tarini Sai Padmanabhuni says the API and MCP service checks audio in 24 ms; a benchmark found 94.47% accuracy and an 8.4% false-positive rate. By Ryan Merket https://runtimewire.com/author/ryan-merket · Published Primary source: X https://x.com/tarini axory/status/2112183048796250209 Why it matters DetectifAI's reported speed may ease integration, while its benchmark flagged 8.4% of genuine clips. Developers evaluating voice-security tools need to weigh latency against false alarms and overall accuracy. DetectifAI https://detectif.ai/ opened its voice deepfake detector to developers on October 19th, offering access through an API and an MCP server. Founder Tarini Sai Padmanabhuni @tarini axory https://x.com/tarini axory advertised a 24-millisecond check time and said the system caught 97.3% of cloned voices in an independent benchmark. When you purchase through links in our articles, we may earn a small commission https://runtimewire.com/affiliate-monetization-standards . This doesn’t affect our editorial independence.