Understand how Negentro works - Negentro Docs

The platform at a glance

Request lifecycle

#	Step	Description
01	Receive the request	Your application accepts a user message or task.
02	Identify the scope	Resolve the applicable user, session, tenant, and permissions.
03	Retrieve information	Search the relevant memory and connected knowledge sources.
04	Assemble context	Select and organize useful results for the task and token budget.
05	Generate a response	Pass the assembled context to your chosen model.
06	Persist useful outcomes	Where appropriate, record new facts, preferences, or decisions for future use.

Separate memory from context

Memory is the information your system retains. Context is the information supplied to a model for a particular operation. Keeping these concepts separate makes it easier to retain useful information without repeatedly sending an entire conversation history to the model.

Design principles

Deployment considerations

Your deployment architecture should define the API boundary, credential management, storage configuration, network access, retention policy, backup and recovery strategy, and observability requirements.