What is the Napster Omniagent API?
The Napster Omniagent API is a full agent platform that lets you create, manage, and deploy AI agents with real-time video avatars, voice, and telephony — all from a single integration. The API is available in public preview on the Azure Marketplace — anyone can sign up and start building today. Every agent you build has five core capabilities: a persistent identity and persona, a domain knowledge base, cross-session memory, real-world tool actions, and built-in safety governance.
What can I build?
- Video AI agents — Lifelike 2D avatars with lip-sync and natural expressions embedded in your web app via WebRTC.
- Voice assistants — Audio-only agents for lightweight integrations using WebSocket connections.
- Telephony bots — Phone-line agents via SIP (live). VoIP support is currently in progress and coming soon.
- Intelligent support agents — Agents backed by your documents, FAQs, and real-world tool integrations.
Tip: The Napster Omniagent API is available on the Azure Marketplace. Test any agent in the Playground before writing a single line of integration code. Full setup guide at developers.napster.com/docs/guides/azure-resource.
Edge MCP — making your website operable by an AI agent
Beyond answering questions about your website, Napster agents can now operate your website using Edge MCP. Your site declares which of its real operations an agent may run — search products, add to cart, place an order, navigate between pages — and the agent calls them during live conversations.
Edge MCP is built on the WebMCP standard, so what you expose works with any compatible agent, including Napster's Omniagent. It is designed to work alongside the Web SDK: the Web SDK puts the agent on your page, Edge MCP gives the agent the ability to act on the page.
This is the first release of Edge MCP. Napster is actively looking for feedback to shape where it goes next. See the Edge MCP docs to get started.