According to a recent LinkedIn post from LiveKit, the company is showcasing AssemblyAI’s Universal-3 Pro speech-to-text model running on LiveKit Inference and integrated into a LiveKit Agent. The post describes four streaming tests focused on entity accuracy, multilingual code-switching, domain-specific vocabulary, and verbatim disfluency capture.
Claim 30% Off TipRanks
- Unlock hedge fund-level data and powerful investing tools for smarter, sharper decisions
- Discover top-performing stock ideas and upgrade to a portfolio of market leaders with Smart Investor Picks
The LinkedIn post highlights that Universal-3 Pro, when accessed via LiveKit Inference, appears to improve handling of structured data such as credit card numbers, emails, and phone numbers versus a prior Universal Streaming model. It also suggests better performance on mixed English-Spanish audio, with language detection enabling both languages to appear correctly in transcripts rather than being forced into one.
According to the post, domain-specific vocabulary can be tuned using a keyterms prompt, with AssemblyAI reportedly indicating up to a 45% accuracy improvement for targeted key terms, and the example outputs are logged and compared through LiveKit Cloud’s observability dashboard. The post further notes that prompting can switch the model into a “verbatim” mode for capturing disfluencies, which may be relevant for legal, medical, or coaching use cases.
From an investor perspective, the content suggests LiveKit is positioning its Inference and Cloud observability offerings as a platform for advanced, real-time speech and AI agent applications. Easier access to Universal-3 Pro through a LiveKit API key could support developer adoption, potentially increasing usage-based revenues and strengthening LiveKit’s role in the conversational AI and real-time communications ecosystem.
The emphasis on structured data accuracy, multilingual support, and domain customization points to potential applicability in contact centers, fintech, healthcare, and compliance-heavy industries. If developers and enterprise customers see measurable transcription gains through this integration, LiveKit could enhance its competitive stance among infrastructure providers that support AI-driven voice and video workflows.