Napster Companion API

Napster Companion API lets you create AI agents that see, speak, and act. Connect your Azure OpenAI real-time deployment, give your agent a face and personality, wire up custom tools, and go live in your application with a drop-in Web SDK. No infrastructure changes, no retraining, no video pipeline to build.

Every agent is an Omniagent: a persistent AI entity with a consistent identity, personality, and behavior across every session and channel. You define who the agent is, what it can do, and how it interacts, then deploy it in your own application through WebRTC (browser) or WebSocket (server-side integrations).

Key Benefits

Lifelike Video Avatars - Create a custom avatar from a single photo. No 3D modeling or animation required. Choose from three display modes: Round (face-focused circle), Rectangle (full body with background), or Silhouette (background removed, floating on your UI). Or pick from the Napster Library of ready-to-use avatars.

Rich, Distinct Personalities - Every agent has a Large Persona Model: over 30 individually scored traits across Big Five personality dimensions, conversational style, and operational behavior. Two agents with the same knowledge will respond differently in tone, pacing, empathy, and formality. Your agents feel like real people, not chatbots.

Custom Tool Execution - Agents don't just talk. They act. Define custom tools that call your APIs, trigger workflows, look up data, or interact with external systems during a live session. Choose implicit execution (your app handles it) or explicit execution (forwarded to your HTTP or WebSocket endpoint). WebSocket-based tools stay connected for the entire session, with optional real-time conversation streaming to your server.

Seamless Azure OpenAI Integration - Connect directly to your existing Azure OpenAI real-time deployment. You provide your endpoint, deployment name, and key. All AI inference runs in your Azure tenant. Napster handles video rendering, agent orchestration, and session management.

Developer-Friendly Web SDK - Integrate using the drop-in Web SDK. Handles WebRTC connection setup, audio-video synchronization, session lifecycle, and provides UI primitives. Supports React, Vue, Angular, vanilla JavaScript, or a simple script tag. No build tools required.

Pay-As-You-Go Pricing - $0.01/minute when you bring your own Azure OpenAI deployment. No setup fees, no commitments. Cancel anytime through the Azure Portal.

How It Works

1. Create your agent - Sign in to the Companion API Dashboard. Create a Companion with a name, personality description, and photo. The system generates a video avatar and configures the agent's persona automatically.
2. Add tools - Define custom tools that connect to your systems. Your agent can call APIs, look up records, and execute real tasks during conversations.
3. Test in the Playground - Assemble your agent by selecting a Companion and attaching tools. Start a live video and voice session to test behavior. Iterate on the system prompt, swap tools, and refine until it's right.
4. Integrate with the Web SDK - Generate an API key, create an auth token on your backend, and initialize the SDK in your app. Your agent is live.

Use Cases

• Customer Engagement - Enhance support, onboarding, or account management with human-like video agents that can look up accounts, create tickets, and take action during the conversation.
• Education & Training - Deliver interactive digital instructors with distinct personalities that respond naturally to learners and adapt their communication style.
• Sales & Marketing - Add conversational spokespeople to product demos and campaigns. Agents can answer questions, qualify leads, and trigger workflows in real time.
• Enterprise Applications - Humanize internal tools, HR assistants, or IT help desks with expressive video agents that execute real tasks, not just talk about them.