📘 A Complete Guide to Understanding the Namirasoft Inference Console

Guide Overview

This guide is designed to help you navigate and fully utilize the Namirasoft Inference Console. Whether you are setting up your first Model or combining inferencers into complete execution flows, this guide provides clear, section-by-section explanations to support confident and informed use of the system.

Introduction to Namirasoft Inference

Namirasoft Inference is where you select the AI models you want to use and define how each request is executed, from routing and fallback to validation, guard rails, and cost limits. Any application can use the models and flows you define here through the Namirasoft Inference API. The console is where you create, configure, and manage every building block described below. Namirasoft applications such as Namirasoft Expert and Namirasoft Job Arranger also use it.

Console Guide Sections

The sections below follow the same structure as the console menu, so what you see here matches what you see when you work. Each entry links to a dedicated guide covering that part of the console in full detail.

Inferencers

Inferencers are the units that answer requests. Every inferencer type below can stand in for any other, so you can combine and nest them freely.

🔹 Model

What’s included: Learn how to create and configure a Model, select the AI model it uses, and choose how that model is accessed.

🔹 Load Balancer

What’s included: Learn how to spread requests across several models using per target weights and an algorithm.

🔹 Failover

What’s included: Learn how to try models in order and continue with the next when one does not respond in time.

🔹 Race

What’s included: Learn how to send a request to several models at once and use the fastest response.

🔹 Router

What’s included: Learn how to send each request down a different path using routes, conditions, and a default inferencer.

🔹 Gateway

What’s included: Learn how to wrap an inferencer with guard rails and knowledgebases so every request is screened before it runs.

Rules

Rules decide how requests are evaluated, so your flows can react to what each request contains.

🔹 Condition

What’s included: Learn how to define a single check on a request, such as contained words, length, tags, or a Classifier result.

🔹 Condition Group

What’s included: Learn how to combine several conditions with And or Or logic into one reusable decision.

🔹 Classifier

What’s included: Learn how to label incoming requests with your own classes and use those labels in routing decisions.

Context

Context is the stored data that surrounds execution: conversation history, reference knowledge, and automatic cleanup.

🔹 Memory

What’s included: Learn how to carry conversation context across chats, independent of the model that answers.

🔹 Knowledgebase

What’s included: Learn how to store items of reference content and flow them into an inferencer as context.

🔹 Wiper

What’s included: Learn how to clean up stored data automatically based on a time or count policy.

Safeguards

Safeguards protect what goes into your models and what comes out of them.

🔹 Validator

What’s included: Learn how to enforce a response format and choose what happens when a response does not match it.

🔹 Guard Rail

What’s included: Learn how to screen user prompts and AI outputs with rules for topics, patterns, and custom prompts.

Spending

Spending keeps AI costs inside the limits you define.

🔹 Budget

What’s included: Learn how to set spending ceilings per run, chat, day, week, and month and share them across inferencers.