Frontend SDK
Best for teams that want the fastest launch. New Port handles the voice-to-avatar pipeline, while your backend only exchanges a short-lived session token for the frontend.
Built for ultra-low latency. We help you deploy photorealistic AI avatars on demand for natural, fluid, real-time conversations with your global audience.
Achieve fluid, sub-100ms conversational latency. Engineered for enterprise production, the engine smoothly scales to support hundreds of concurrent video sessions without hitting hardware bottlenecks.

Connect seamlessly to leading LLMs or directly mount your proprietary enterprise knowledge base. Empower your avatars to deliver accurate, specialized, real-time answers based entirely on your business logic.

Start with a few lines of frontend SDK, or connect Live Avatar to your own agent, voice stack, and RTC infrastructure. Each path gives developers a clear tradeoff between launch speed, control, and deployment flexibility.
Best for teams that want the fastest launch. New Port handles the voice-to-avatar pipeline, while your backend only exchanges a short-lived session token for the frontend.
Best for serverless or private-network environments. Keep full control of LLM, Agent, business rules, ASR or TTS choices, while New Port manages RTC streaming and avatar rendering.
Best for teams building a custom real-time voice agent. Your agent joins New Port's RTC room, subscribes to user audio, runs inference, and publishes audio that drives the avatar.
Best for private deployment or existing RTC infrastructure. You own the SFU, room, and token flow; New Port joins as a rendering layer that turns your agent audio into a live avatar.
Proven deployment scenarios across industries

Deploy proactive AI sales agents on your landing pages to communicate face-to-face with prospects, handle objections, and qualify leads automatically.
Program avatars with specific personas and objections for realistic sales or compliance roleplay, accelerating your internal Learning & Development efficiency.

Replace static chatbots with empathetic video agents capable of resolving complex customer inquiries in real time.
A single custom AI avatar switches across 32 languages in real time, adapting to each user's preference without reloading or handoff.
Deploy highly interactive avatars for HR onboarding or patient intake, providing a warm, guided experience while maintaining strict privacy and data security.


Embed our rendering engine into hotel concierges, airport wayfinding screens, or retail kiosks, transforming static displays into interactive digital humans.
New Port Labs products are available to select enterprise partners before general release. Apply to get hands-on access and shape the product roadmap alongside our research team.