cassetteai.com

What It Is

CassetteAI positions itself as “the sound layer for software.” Built by a small research team in Salt Lake City, it combines audio DSP and diffusion research into a generative audio engine. It covers three modalities: music, sound effects, and voice. Its core value proposition is not just web-based creation, but embedding generated audio directly into games, calls, browser tabs, or applications.

Core Capabilities

The most notable claims are real-time performance and on-device execution: CassetteAI says its streaming response time is under 50 milliseconds, and that its model can run inside an app bundle, reducing server-side latency and privacy overhead. Its SFX Generator can create sound effects up to 30 seconds long, with around 1 second of processing time. The product also offers an SDK and plans to provide, or already provides, a hosted API for developers who cannot deploy on-device. Note that the page does not disclose specific model architecture, training data, supported platforms, SDK languages, or quality evaluation metrics.

Pricing and Trial

Pricing is relatively clear: sound effect generation costs $0.01 per generation, while music costs $0.02 per output minute. This is a pay-as-you-go model, suitable for small-scale initial integration and scaling by playback volume. The captured content does not mention a free quota, trial, voice generation pricing, enterprise plans, SLA, or payment methods.

Pros and Cons

The strengths are its focused use cases, clear latency target, on-device inference that benefits interactivity and privacy, and coverage of three common production needs: music, sound effects, and voice. The downside is that public information remains limited: Chinese-language support, copyright licensing, data retention, compliance certifications, platform compatibility, and customer support are not explained. The timeline also includes future milestones, so the actual availability of features still needs further verification.

Who It’s For and Access from China

CassetteAI is better suited to development teams building games, creator tools, accessibility products, robotics, and other products that need “real-time sound feedback,” rather than casual users who simply want to generate a single song. The captured text provides no information on access from China, so it is currently unknown; payment methods are also not disclosed. For domestic Chinese teams seeking alternatives, options to compare include ElevenLabs, Suno, Stable Audio, and AudioCraft, but network accessibility, commercial licensing, and Chinese-language capabilities should be checked individually.