Expand your ideas with the wide variety of nodes provided by Agentria.

Refer to the descriptions by category to select the right nodes for your workflow and build better abilities.

AI Nodes

✨ Azure OpenAI LLM

A node that provides access to Azure OpenAI’s Large Language Models(LLMs).

It can perform a variety of natural language tasks such as text generation, question answering, summarization, and translation.

✨ Bedrock Anthropic LLM

A node that provides Anthropic’s LLMs via AWS Bedrock.

It emphasizes safety and contextual understanding, making it suitable for conversational responses and analytical tasks.

✨ Google Gemini LLM

A node that provides Google’s Gemini LLMs.

Supports multimodal processing (text + images, etc.), enabling more advanced AI workflows.

✨ Azure Text Embedding

A node that uses Azure OpenAI embedding models to convert text into vectors (numerical representations).

These vectors can be used for semantic search, document similarity, recommendation systems, and other AI-driven tasks.

✨ Anthropic LLM

A node that provides Anthropic’s LLMs.

It specializes in generating AI responses with strong emphasis on ethical guidelines and safety.

✨ LLM Reserve

A backup node that automatically switches to an alternative LLM when the primary model fails or is unavailable.

Ensures service reliability by preparing a reserve language model.

✨ OpenAI LLM

A node that uses OpenAI’s Large Language Models (LLMs).

It can be applied to wide range of natural language tasks, including text generation, summarization, and question answering.

AI Agent Node

💼 Short-term Memory Reader

A node that provides short-term memory functionality by storing and retrieving recent conversation context within a session.

It helps maintain conversational consistency and ensures that user inputs are processed in continuity. This is especially useful for building chatbots or conversational agents that need to remember what was said a few turns ago.

• Scope: Limited to the current session (resets after session ends)

• Use Cases: Carrying on small talk, remembering the last command, maintaining context in interactive flows

💼 Agent Tools

A node that analyzes the user’s message and determines which ability to execute automatically.

Helps agents select and trigger the right functions when multiple abilities are available.

💼 Long-term Memory Reader

A node that provides long-term memory functionality by saving and retrieving organized conversation history and experiences across sessions.

It enables agents to recall user preferences, profiles, and past interactions, allowing for highly personalized responses and recommendations.

• Scope: Persists across multiple sessions (until explicitly updated or deleted)

• Use Cases: Remembering a user’s name, learning their favorite settings, tailoring responses based on historical behavior

💼 ReAct Agent

A node that implements the ReAct (Reasoning + Acting) framework.

It allows LLMs to reason, take actions, and observe outcomes iteratively, making them capable of solving complex tasks with higher accuracy and reliability.

Tools 노드

🛠️ Notion

A node that lets an agent read, write, or manage content within a Notion workspace.

🛠️ Gmail

A node that allows an agent to read and process emails from a Gmail account.

🛠️ Discord

A node that allows an agent to interact with Discord channels, messages, and user events.

Web Tools

🌐 Web Request

A node that sends HTTP requests to external web services and processes the responses inside a workflow.

Can be used for REST API calls, data collection, or webhook handling.

🌐 SerpApi Web Search

A node that performs web searches through the SerpApi service.

It can also be used with DuckDuckGo search results via SerpApi integration.

🌐 DuckDuckGo Web Search

A node that retrieves structured web search results instantly via the SerpApi service.

It allows you to conveniently receive search engine results—such as Google, Bing, or DuckDuckGo—in JSON format.

🌐 Exa Web Search

A node that retrieves high-quality, relevant web search results using the Exa API.

Exa is designed for LLM-based applications, optimized for processing natural language queries and delivering refined information.

File Handling Tools

📎 PDF Text Reader

A node that extracts and reads text content from PDF files.

It uses a PDF.js-based extraction library to parse structured text within documents.

📎 Upstage AI OCR

A node that provides OCR (Optical Character Recognition) functionality to extract text from images or scanned documents.

Upstage’s Document AI is specialized in high-accuracy recognition for Korean and English, making it useful for parsing receipts, scanned forms, and image-based documents.

Code

💻 Python

A node that lets you write and execute Python code directly.

It can also call other nodes like functions within the workflow, supporting tasks such as data preprocessing, computation, or API calls.

Database

🗄️ Embedding Data Searcher

A node that searches stored embedding (vector) data.

Useful for semantic queries, RAG-based document retrieval, or recommendation systems.

🗄️ Structured Data Reader

A node that reads structured data from storage. Supports formats such as tables, database records, or JSON.

🗄️ Structured Data Writer

A node that writes structured data into storage.

Can be used to insert new records or structured datasets for retrieval-augmented applications.

🗄️ Embedding Data Writer

A node that converts text into embedding vectors and stores them.

Useful for later semantic search or context retrieval by saving documents or conversation logs in vectorized form.

* RAG (Retrieval-Augmented Generation) : A method that combines retrieval and generation to improve accuracy.