Glosario

A

API

A set of endpoints that allows applications to communicate with NovaAI services.

API Key

A secret credential used to authenticate requests to the NovaAI API.

Agent

An AI-powered assistant capable of reasoning, using tools, and completing multi-step tasks.

Authentication

The process of verifying API requests using a valid API key.

B

Batch Request

A single request that processes multiple operations together for improved efficiency.

Billing

The system used to calculate and manage API usage charges.

Bearer Token

A token included in the Authorization header to authenticate requests.

C

Chat Completion

An API endpoint for generating conversational responses using language models.

Context Window

The maximum number of tokens a model can process in a single request.

Completion

The output generated by a language model in response to user input.

D

Dataset

A collection of training examples used for fine-tuning models.

Deployment

The process of releasing an AI application into a production environment.

E

Embeddings

Numerical vector representations of text used for semantic search and retrieval.

Endpoint

A specific URL that provides access to a NovaAI API service.

F

Fine Tuning

The process of adapting a model using custom training data.

Function Calling

A capability that allows models to invoke external tools or APIs.

G

Generation

The process of creating text, images, audio, or other AI-generated content.

GPU

Graphics Processing Unit used to accelerate AI model inference and training.

H

HTTP Request

A request sent from an application to a NovaAI API endpoint.

HTTPS

A secure protocol that encrypts communication between clients and servers.

I

Image Generation

Creating original images from natural language prompts using AI models.

Inference

The process of generating predictions or responses from a trained model.

J

JSON

A lightweight data format commonly used for API requests and responses.

K

Knowledge Base

A collection of documents used by AI agents and RAG applications.

L

Latency

The time required for an API request to receive a response.

LLM

Large Language Model capable of understanding and generating natural language.

M

Message

A conversation object containing a role and content within chat requests.

Model

An AI system that processes input and generates intelligent outputs.

Multimodal AI

AI capable of understanding multiple input types such as text, images, and audio.

N

Node.js

A JavaScript runtime commonly used with the NovaAI SDK.

NovaAI SDK

Official libraries that simplify integration with NovaAI APIs.

O

Output Tokens

Tokens generated by the model as part of its response.

P

Prompt

Instructions or questions provided to a model to generate a response.

Project

An isolated workspace for managing API keys, usage, and resources.

Q

Query

A search request or user input sent to a model or vector database.

Queue

A temporary waiting state before a request or training job begins processing.

R

Rate Limit

The maximum number of requests allowed within a specific time period.

Response

The structured output returned by the NovaAI API.

Retrieval-Augmented Generation (RAG)

A technique that combines document retrieval with language model generation.

S

SDK

Software Development Kit for integrating NovaAI into applications.

Streaming

Receiving model output incrementally as it is generated.

System Message

Instructions that define the assistant's behavior during a conversation.

T

Temperature

A parameter that controls the randomness and creativity of model responses.

Token

The basic unit used by models to process text.

Tool

An external capability that an agent can use to perform actions.

U

Usage

The amount of resources consumed by API requests.

User Message

Input submitted by the end user during a conversation.

V

Vector

A numerical representation of text generated by an embedding model.

Vector Store

A database optimized for storing and searching vector embeddings.

Vision

A multimodal capability that enables AI models to understand images.

W

Webhook

An automated HTTP callback triggered by specific application events.

Workspace

A shared environment for managing projects, team members, and billing.

X

XML

A structured markup format occasionally used when exchanging data with external systems.

Y

YAML

A human-readable configuration format commonly used in development workflows.

Z

Zero-Shot Prompting

A prompting technique where the model completes a task without prior examples.