Create a New API Key
Generate an API key from the Napster dashboard to authenticate with the Napster API
API keys authenticate your requests to the Napster API. Each key is scoped to a project, so you can manage access separately for different environments (e.g. staging vs production). A key also sets the architecture your Omniagent runs on and the per-minute rate for its sessions.
Your API key is a secret. Keep it private and only use it on the server side — never expose it in client-side code.
Navigate to Keys
Open the Napster dashboard and select your organization and project from the top navigation. Then click Keys in the left sidebar under Build.
Create a new API key
Click the + Create API key button in the top right corner.
Configure the key
Fill in the form:
| Field | Description |
|---|---|
| Key Name | A descriptive label (e.g. Stage API Key, Prod API Key) |
| Project | Automatically set to your current project |
| Architecture | How your Omniagent listens, thinks, and speaks: Realtime, Cascade, or Other (see below). Each architecture has its own provider or platform options. |
The Pricing panel beside the form shows the per-minute rate for your current selection. You're only charged while an agent is running.
Architectures
One model listens, thinks, and answers in a single step, so conversations feel quick and natural. Realtime is the default and the simplest way to get a good result.
Choose a Provider:
| Provider | Description |
|---|---|
| Azure OpenAI | Default provider. Works with all channels including phone. Runs on Napster's managed infrastructure — no credentials needed. Tick Use my own credentials to connect your own Azure OpenAI deployment instead (deployment name, endpoint URL, API key, and optionally a model). |
| OpenAI | Connects directly to OpenAI. Always requires your own OpenAI API key (and optionally a model). Required for MCP connectors. |
Realtime sessions cost $0.0583 per minute on Napster's managed infrastructure, or $0.01 per minute when you use your own credentials.
Speech recognition, the language model, and the voice run as separate steps instead of one realtime model, and you choose the provider for each one. Use Cascade when you want control over the individual parts.
| Stage | What it does | Providers |
|---|---|---|
| ASR | Turns what the user says into text | Azure OpenAI |
| LLM | Takes what the user said and decides what to reply | Azure OpenAI, Anthropic |
| TTS | Speaks the reply out loud | Azure OpenAI, OpenAI |
Azure OpenAI stages run on Napster's managed infrastructure with a preset model. Tick Use my own credentials on a stage to connect your own Azure OpenAI deployment for that stage. Anthropic and OpenAI stages always require your own API key and a model name.
Some features work differently on Cascade — see Cascade architecture.
Your agent already exists on another platform. The Napster API connects to it and gives it a face — a lifelike video avatar on top of your existing agent logic.
Humain and Microsoft Foundry are not enabled by default. Contact us to request access for your organization.
Choose a Platform:
| Platform | Description | Required fields |
|---|---|---|
| Humain | Powered by Allam, an Arabic-first foundation model built by HUMAIN in Saudi Arabia with strong dialect coverage and cultural alignment. Humain provides speech recognition, the model, and the voice as one package. Best for Arabic-language agents. | API URL, API path, API key, LLM base URL, LLM API key, LLM model |
| Microsoft Foundry | Connects to an agent you've built on Microsoft Foundry. | Foundry project endpoint, agent name, agent version, tenant ID |
For Microsoft Foundry, here's where to find each value:
| Field | Where to find it |
|---|---|
| Foundry project endpoint | Your project's Home page in Microsoft Foundry, e.g. https://<resource>.services.ai.azure.com/api/projects/<project> |
| Agent name | The Agents page of your project |
| Agent version | The Version column on the Agents page, e.g. 1 |
| Tenant ID | Microsoft Entra ID → Overview in the Azure Portal |
Before you use a Foundry key, give Napster access to your Foundry resource and make sure your agent has Voice mode on. See Connect a Microsoft Foundry agent.
Click + Create API Key to generate the key.
Copy the key
The API Key Generated page shows your new key. Copy it from the API key field, then click ← Back to keys. The key now appears in your keys list with its provider, deployment, and rate.
Store the key
Save the key as an environment variable in a gitignored .env file:
NAPSTER_API_KEY=your-api-key-hereAll API requests use the X-Api-Key header for authentication:
curl https://companion-api.napster.com/public/companions/napster-stock?pageSize=1 \
-H "X-Api-Key: $NAPSTER_API_KEY"A 200 response confirms your key is working.
Next steps
- Authentication — full details on how authentication works
- Cascade architecture — which features behave differently on a Cascade key
- Connect a Microsoft Foundry agent — set up access and create a Foundry key
- Quickstart — build and deploy your first Omniagent