Napster Companion API

Napster Companion API is the platform for building Omniagents: persistent AI agents with a consistent identity, personality, knowledge, and behavior across every channel. Define the agent once, and the same agent shows up across video, audio, and telephony. Same face, same voice, same memory, same tools, regardless of how users connect.

Connect your own OpenAI real-time deployment on Microsoft Foundry, or use the Napster Hosted tier with a managed LLM included. Give your agent a face and a personality, plug in your knowledge and tools, and go live with a drop-in Web SDK. No infrastructure changes, no retraining, no video pipeline to build.

One agent, every channel

Video - Real-time avatars with synchronized lip-sync and expressive gestures. Embed via the Web SDK over WebRTC.

Audio - Voice-only agents for audio-first experiences. Low bandwidth, low latency, over WebSockets.

Telephony - Connect your Omniagent to phone lines via VoIP and SIP. Same intelligence, knowledge, and memory as your web sessions.

Key Benefits

Lifelike Video Avatars - Create a custom avatar from a single photo. No 3D modeling required. Three display modes (Round, Rectangle, Silhouette), or pick from thousands of pre-built avatars.

Rich, Distinct Personalities - Every Omniagent has a Large Persona Model: 30+ scored traits across Big Five dimensions, conversational style, and behavior. Your agents feel like real people, not chatbots.

Knowledge and Memory - Upload documents, websites, and Q&A pairs. Agents retrieve the right context at conversation time and remember context across sessions and channels.

Custom Tool Execution - Agents don't just talk. They act. Define tools that call your APIs, trigger workflows, or look up data during a live session. Implicit or explicit execution, with optional real-time conversation streaming to your server.

Bring Your Own LLM - Connect your OpenAI real-time deployment on Microsoft Foundry. Inference runs in your Azure tenant. Your models, your data, your control. Or use the Napster Hosted tier.

Safety and Governance - Content moderation, crisis detection, and full audit trails. SOC 2 Type II certified, GDPR and CCPA compliant, AES-256 encryption at rest.

Drop-in Web SDK - Handles WebRTC setup, audio-video sync, and session lifecycle. Works with React, Vue, Angular, or vanilla JS.

Pay-As-You-Go Pricing - $0.01/min with your own model deployment, or $0.058/min on the Napster Hosted tier. No setup fees, no commitments.

Use Cases

• Customer Engagement - The same Omniagent greets customers on your website and takes phone calls. It looks up accounts, creates tickets, and takes action with full memory of past interactions.
• Education and Training - Interactive digital instructors with distinct personalities that respond naturally to learners across video and audio.
• Sales and Marketing - Conversational spokespeople for demos and campaigns. Agents qualify leads, answer questions, and trigger workflows in real time.
• Manufacturing - The conversational layer of the factory. Operators, supervisors, and engineers talk to a single Omniagent on the shop floor, in the control room, or over voice. Memory grows from those conversations and from ambient factory events streamed directly into the agent, so it always knows what's happening on the line and can answer, advise, and act in context.
• Enterprise Applications - Humanize internal tools, HR assistants, and IT help desks with expressive agents that execute real tasks.