> ## Documentation Index
> Fetch the complete documentation index at: https://docs.lyzr.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Glossary

> Key terminology and definitions for Lyzr Agent Studio and generative AI concepts.

### Agent

An autonomous AI component configured to perform tasks, answer queries, or orchestrate workflows using Lyzr’s framework.

### API (Application Programming Interface)

A set of RESTful endpoints that allow programmatic interaction with Lyzr agents, tools, and services.

### AIMS (AI Management System)

Lyzr’s console for monitoring, managing, and governing agent deployments, usage metrics, and audit logs.

### Audit Log

A chronological record of events and actions taken by agents, users, or systems—used for compliance and debugging.

### Bias Mitigation

Techniques and processes to detect and reduce unwanted biases in AI outputs, ensuring fair and equitable responses.

### Chain of Thought

A reasoning technique where intermediate steps are made explicit in prompts to improve model inference and explainability.

### Connector

A pre-built integration between Lyzr agents and external services (e.g., Slack, Salesforce, Google Sheets) to perform actions or fetch data.

### DAG Orchestration

Directed Acyclic Graph–based workflow execution where tasks run in sequence or parallel based on explicit dependencies.

### Data Ingestion

The process of importing documents, databases, or other data sources into Lyzr for indexing and retrieval.

### Embedding

A numeric vector representation of text (or other data) capturing semantic meaning for similarity comparisons and search.

### Embedding Model

An algorithm that converts input data (e.g., text, images) into embeddings; common examples include OpenAI or Cohere embedding models.

### Fine-Tuning

The process of adapting a pre-trained LLM to a specific task or domain by training on a smaller, task-focused dataset.

### Generative AI

AI systems capable of producing new content—text, images, audio—based on learned patterns from training data.

### Intent Recognition

The process by which an AI agent identifies a user’s goal or purpose from their input to select appropriate actions.

### JSON Schema

A JSON-based format for defining the structure, required properties, and data types of JSON documents or API payloads.

### Knowledge Base

A repository of documents, web pages, or databases that agents query to ground responses in real-world information.

### LLM (Large Language Model)

An advanced neural network model (e.g., GPT, Gemini, Claude) trained to understand and generate human-like text at scale.

### Managerial Orchestration

A dynamic orchestration mode where a “manager” agent decomposes objectives into subtasks and dispatches worker agents at runtime.

### Memory Modules

**Short-Term Memory:** Stores context for the duration of a single conversation.
**Long-Term Memory:** Persists context across sessions to personalize interactions over time.

### Model-Agnostic Pipeline

Lyzr’s architecture allowing you to swap between different LLMs (OpenAI GPT, Gemini, Claude, Bedrock, etc.) without rebuilding workflows.

### Multi-Agent Workflow

An automated process in which multiple specialized agents interact or run in sequence to achieve complex tasks.

### Orchestration

The coordination of multiple steps, tools, or agents into a single automated workflow to achieve an end-to-end process.

### Parser

A component that processes and structures raw data (e.g., JSON, HTML) into a format suitable for downstream tasks or prompts.

### Prompt

A piece of text given to an LLM that guides it to produce a desired output, such as questions, instructions, or context.

### Prompt Template

A reusable blueprint for prompts with placeholders that can be filled dynamically at runtime with variables or data.

### RAG (Retrieval-Augmented Generation)

A hybrid approach combining information retrieval from external sources with generative models to produce grounded responses.

### Reranker

A module that reorders retrieved results (documents, passages) based on relevance before feeding them into a generative model.

### Role-Based Access Control (RBAC)

A security paradigm where permissions are assigned to roles rather than individuals, simplifying governance and compliance.

### SDK (Software Development Kit)

A set of libraries, tools, and documentation that simplifies integration with Lyzr from programming environments like JavaScript or Python.

### Studio

Lyzr’s no-code visual interface for designing, testing, and deploying agents without writing any code.

### Tokenization

The process of breaking text into smaller units (tokens) that a language model can process efficiently.

### Vector Store

A specialized database optimized for storing and querying vector embeddings for fast semantic search and similarity matching.

### Webhook

A user-defined HTTP callback URL that receives real-time notifications or data from Lyzr agents when specific events occur.

### Zero-Shot

An inference mode where an LLM performs a task without any examples or prior fine-tuning, relying solely on its pre-trained knowledge.

### Few-Shot

An inference mode where an LLM is provided with a small number of examples in the prompt to guide its output for a specific task.

### KPI (Key Performance Indicator)

A measurable value used to evaluate the success of an agent or workflow against business objectives (e.g., accuracy, response time).