Skip to main content
Use a trusted backend to keep Neuro credentials out of browser and mobile code. Start with the tested backend sample, then add the capability your application needs.

Architecture

The application backend protects Agent API secrets and coordinates persistence. A browser can participate in the dedicated web-form onboarding and session flow; it must not receive a trusted account-creation secret.

Prerequisites

  • A confirmed development environment.
  • A supported Node.js runtime for the current Agent API Quickstart.
  • The host and credentials required by that Quickstart.
  • A storage mechanism for application state and Neuro identifiers.

Build sequence

  1. Run the Agent API Quickstart.
  2. Run the trusted-backend sample from a clean repository clone.
  3. Store environment configuration outside the source repository.
  4. Confirm that only allowlisted account fields reach the browser.
  5. Add an identity, agreement, asset, payment, messaging, or AI-tool capability that your Neuron supports.
  6. Store the identifiers and state named by that capability’s guide.
  7. Test the success, rejected-session, invalid-configuration, and unavailable-Neuron paths supported by the environment.

Expected result

The client shows account information obtained through the application backend. No Agent API secret, password, JWT, or private key appears in the browser source, URL, analytics, or logs.

State created

Next step

Choose a capability from What can I build?. Each capability page tells you what you can implement now and what you must confirm with the Neuron operator.