Agents

Agents

Neutron AI helps agents remember useful context across sessions without turning your product into an unbounded chat log. Each agent should have a clear purpose, explicit scopes, and predictable memory behavior.

Define these fields before connecting an agent to Neutron:

FieldPurpose
agentIdStable identifier for the workflow or assistant.
PurposeWhat the agent is allowed to help with.
Input contractWhat the user or system can provide.
Output contractWhat the agent must return.
Scope policyWhich tenant and scope IDs can be used.
Memory policyWhat can be remembered, recalled, reflected, or forgotten.
Safety policyWhat should never be stored or revealed.

Agent Loop

flowchart LR
  A["User task"] --> B["Request scoped context"]
  B --> C["Call model provider"]
  C --> D["Return product response"]
  D --> E["Reflect useful outcome"]
  E --> F["Store safe memory"]

Keep the current user task as the primary instruction. Treat Neutron memory as supporting context that can improve continuity and personalization.

Guides