SAN FRANCISCO—AudioShake has launched its first streaming-capable SDKs designed specifically for real-time music detection and copyright-safe audio cleanup. The SDK enables broadcasters, sports leagues, and media creators to automatically detect and remove copyrighted music from live and post-production content—helping them avoid takedowns, fines, muted audio, and distribution delays.
“Our partners are using AudioShake’s tools not just for compliance, but to reimagine what’s possible with their existing content,” said Jessica Powell, CEO of AudioShake. “We’ve helped teams clean up archival footage, prep assets for international distribution, and meet the demands of social content at scale. Now, with real-time capabilities, we’re unlocking entirely new use cases—from mobile apps and on-the-fly editing, to live broadcast cleanup and programmatic compliance checks. And thanks to an architecture that is optimized for NVIDIA GPUs, the SDK runs at blazing fast speeds.”
Background music has become one of the biggest and fastest-growing copyright liabilities in live and social video. Sports teams, creators, and broadcasters routinely capture licensed songs in arenas, locker rooms, events, and public spaces—triggering copyright claims that can result in financial penalties, blocked global feeds, or fully muted audio.
AudioShake said that the new SDKs address this challenge by pairing industry-leading sound separation technology with low-latency inference. The system identifies music in real-time—even in noisy environments like stadiums or live events—and automatically removes it while preserving dialogue, crowd noise, effects, and other ambient sound without interrupting workflows.
Beyond copyright compliance, the SDK provides fine-grained control over dialogue and background sound, enabling creative teams to quickly rebalance noisy mixes, boost speech, or reduce crowd roar during fast-paced production, the company explained.
Audioshake is a member of the NVIDIA Inception program for startups, and leverages NVIDIA accelerated computing and NVIDIA Dynamo-Triton for training and cloud-based inference. Its real-time SDKs utilize NVIDIA accelerated computing directly via optimized runtime frameworks on Linux and Windows to deliver low-latency, high-accuracy performance in dynamic production environments.
Major media companies—including most large US film studios and broadcasters, as well as BET, NFL Films, TVPuls, and ANT1 TV—are already using AudioShake’s technology across workflows spanning dialogue isolation, music removal, and audio cleanup, the company reported.
"As demand grows for real-time, AI-powered audio workflows, advanced technologies are critical to meet industry needs and keep up with the rapidly evolving landscape of sound production, editing, and processing," said Shari Reich, developer relations, audio and music at NVIDIA. "With NVIDIA accelerated computing and software, AudioShake’s solution empowers content creators, broadcasters and developers to perform high-quality audio separation in real-time to enable seamless integration into live broadcast environments."
The AudioShake SDK is now available for iOS/macOS, Android, Windows, and Linux. More information is available at developer.audioshake.ai.
