Agentifact assessment — independently scored, not sponsored. Last verified Mar 6, 2026.
Nuance (Microsoft)
Nuance, now part of Microsoft, provides enterprise voice AI solutions including Dragon speech recognition, the Nuance Healthcare Developer Platform, and the Azure-hosted Voice Live API for low-latency speech-to-speech voice agents. The Voice Live API (released mid-2025) unifies speech recognition, generative AI, and TTS into a single real-time interface with tiered Lite, Basic, and Pro tiers based on the underlying generative AI model. Dragon Medical One and related products target clinical speech documentation. Pricing varies by product: Azure Speech Services start at competitive per-minute rates; enterprise and healthcare products require custom contracts.
Viable option — review the tradeoffs
You need to build a real-time voice agent that handles speech recognition, conversational AI, and text-to-speech in a single unified pipeline without managing multiple disparate services.
Low-latency speech-to-speech interactions with 600+ voices across 150+ locales. Semantic VAD (checks conversation completeness by word content, not just silence) reduces false turn-endings. Avatar lip-sync available. Performance scales automatically; you pay per minute of interaction. Quirk: API design mirrors Azure OpenAI Realtime API but adds optional Speech-specific features (noise suppression, echo cancellation, advanced turn detection) that don't break existing Realtime API code.
You're building a customer service or healthcare voice bot and need to customize both speech recognition (for domain-specific terminology) and voice output (for brand personality) without rebuilding the entire pipeline.
Custom voices and phrase lists work reliably but require upfront definition. Phrase lists are 'just-in-time' and lightweight; custom speech models take longer to train. Healthcare products (Dragon Medical One) operate on separate licensing and require custom contracts—not included in Voice Live's per-minute pricing.
Healthcare and Dragon products operate outside Voice Live's unified pricing
While Nuance's Dragon Medical One and related clinical products are part of the Microsoft portfolio, they are not integrated into Voice Live API's tiered per-minute pricing model. Healthcare builders must negotiate separate enterprise contracts and manage Dragon as a distinct service, negating some of the 'single pipeline' benefit.
Voice Live API is in public preview (as of May 2025); production readiness varies by region
Voice Live was announced at Microsoft Build 2025 and entered public preview mid-2025. Availability, SLA terms, and feature stability may vary by Azure region. Confirm regional availability and preview status before committing production workloads. Pricing and tier definitions may change before general availability.
Voice Live API is a superset: it mirrors Realtime API's event model but adds speech-specific enhancements (semantic VAD, noise suppression, echo cancellation, avatar, custom voices) and handles speech-to-text natively.
You need end-to-end voice agent capabilities (speech recognition → AI → TTS) with minimal orchestration, or you want avatar output, semantic turn detection, or custom voice branding.
You're already invested in Realtime API and only need text-in/audio-out, or you need maximum flexibility to swap speech and TTS providers independently.
Trust Breakdown
What It Actually Does
Nuance provides AI-powered speech recognition and clinical documentation tools that help healthcare organizations reduce administrative burden and improve patient care, plus contact center solutions for customer service automation across industries.
Nuance, now part of Microsoft, provides enterprise voice AI solutions including Dragon speech recognition, the Nuance Healthcare Developer Platform, and the Azure-hosted Voice Live API for low-latency speech-to-speech voice agents. The Voice Live API (released mid-2025) unifies speech recognition, generative AI, and TTS into a single real-time interface with tiered Lite, Basic, and Pro tiers based on the underlying generative AI model. Dragon Medical One and related products target clinical speech documentation.
Pricing varies by product: Azure Speech Services start at competitive per-minute rates; enterprise and healthcare products require custom contracts.
Score Breakdown
Protocol Support
Capabilities
Governance
- permission-scoping
- audit-log
- rate-limiting
- resource-limits