Overview
What the Companion API is, what you can build with it, and how to get started
The Companion API lets you create, manage, and deploy AI agents with persistent identity, memory, knowledge, real-world actions, and built-in safety — across text, voice, and video.
Most agent platforms give you a stateless chatbot that forgets users between sessions, can't take real actions, and requires a separate setup for every channel. The Companion API gives you a single agent that maintains a consistent persona and memory across every interaction, executes real workflows through tool integrations — all with content moderation, topic restrictions, and auditability built in.
What you can build
Every agent you create through the Companion API has five core capabilities:
Identity
Define your agent's role, persona, voice, and visual appearance. These persist across every session and surface, so your agent feels consistent whether a user reaches it on web, mobile, or phone.
Knowledge
Give your agent domain expertise by ingesting files, FAQs, and structured data. Your agent draws on this knowledge to answer questions accurately instead of relying solely on the base LLM.
Memory
Your agent remembers who it's talking to and what was discussed — across sessions and across channels. When a user returns, the conversation picks up where it left off.
Actions
Connect your agent to real systems through tools and functions so it can trigger workflows and generate content — not just talk.
Governance & Safety
Content moderation, crisis detection, topic restrictions, and full session auditability are always active at the platform level. You configure the boundaries; the platform enforces them.
Capabilities
| Capability | Details |
|---|---|
| Real-time video avatars | 2D avatars with lip sync and natural expressions |
| Persistent identity and memory | Agents maintain a consistent persona and remember users across sessions and channels |
| Tool integration | Connect agents to real systems through custom functions to trigger workflows, generate content, and more |
| Model-agnostic | Bring your own Azure OpenAI Real-Time deployment, or use a Napster-managed one |
| Web SDK | Works with React, Vue, Angular, or vanilla JS; mount a companion in any DOM container |
| Low latency | ~300ms response times |
| Multilingual | ~30 languages supported |
| 2,000+ companions | Pre-built expert agents, or wire your own knowledge base and tools |
| Built-in safety | Content moderation, crisis detection, topic restrictions, and auditability are always active |
Compliance and security
- SOC 2 Type II — report available under NDA
- GDPR & CCPA — US and EU data protection
- Encryption — AES + TLS 1.2+ at rest and in transit
- Pen testing — annual third-party security audits
- Data residency — regional options for enterprise
- Tenant isolation — your data stays in your environment; no cross-customer data mixing, no training on your data