Blocks Glossary

A comprehensive list of all available blocks in RAPFlow with brief descriptions, organized exactly as they appear in the Flow Editor palette. Click on any block name to view detailed documentation.

Common Blocks

inject

Manually trigger flows and inject data into workflows. Essential for starting workflows and testing individual components.

debug

Monitor data flow and debug issues by displaying message contents in the debug panel. Invaluable for troubleshooting workflows.

complete

Mark the completion of workflow execution paths. Used to signal successful completion of processing chains.

catch

Handle errors and exceptions that occur in your flow. Prevents workflows from failing and enables graceful error recovery. Essential for HTTP flows to catch errors and return proper responses to clients.

status

Display and monitor the status of nodes in your workflow. Provides visual feedback on node states.

link in

Create virtual input connections to organize complex flows. Enables clean flow organization without physical wire connections.

link call

Call linked subflows or functions. Enables modular workflow design and reusable components.

link out

Create virtual output connections to organize complex flows. Complements link in blocks for clean flow organization.

comment

Add documentation and notes to your flows. Helps maintain readable and well-documented workflows.

extract_archive

Extract files from compressed archives (ZIP, TAR, etc.). Useful for processing batched documents or uploaded file collections.

python_custom_shell

Execute custom Python scripts within your workflow. Enables integration of Python libraries and custom processing logic. Requires virtual environment to be created using python_env_setup block first. Only one print statement allowed in main.py - output becomes msg.payload.

python_env_setup

Configure Python environment settings and dependencies. Prepares the environment for Python-based processing blocks. Used only once to create virtual environment - doesn't need to run every time during flow execution. Virtual environment names must be unique within the Flow AI project.

Text Processor

Advanced text processing and manipulation capabilities. Handles text cleaning, transformation, and analysis operations.

utils

Utility functions including the ability to add bounding boxes to images. Essential for object detection workflows and image annotation.

Function Blocks

function

Execute custom JavaScript code within your flow. Provides unlimited flexibility for custom logic and data manipulation.

switch

Route messages to different paths based on conditions. Enables conditional logic and branching in workflows. Routes messages to different outputs based on property values, with "Otherwise" output for unmatched messages.

change

Modify message properties, set variables, and transform data structures. Core block for data manipulation.

range

Map input values to different output ranges. Useful for normalizing data and scaling values.

template

Generate formatted text output using templates with variable substitution. Perfect for creating reports and notifications.

delay

Introduce delays in flow execution. Useful for rate limiting, timing coordination, and scheduled operations.

trigger

Advanced triggering and timing control for workflow execution. Provides sophisticated trigger logic beyond basic inject blocks.

filter

Filter messages based on specified criteria. Enables selective processing of data streams.

counter loop

Execute repeated operations with a counter. Ideal for batch processing with iteration tracking.

array loop

Iterate through arrays and process each element individually. Essential for handling collections of data.

while loop

Execute conditional loops that continue until a condition is met. Enables complex iterative processing.

Network Blocks

http in

Receive HTTP requests and trigger workflows via web endpoints. Enables REST API integration and webhook handling. Request body becomes msg.payload, headers in msg.headers, query parameters in msg.query. Default timeout is 8 minutes.

http response

Send HTTP responses back to clients. Essential for building API endpoints and web services.

http request

Make HTTP requests to external services and APIs. Enables integration with third-party systems and data sources.

Sequence Blocks

split

Split strings into arrays or break apart data structures. Useful for parsing delimited data and text processing.

join

Combine arrays into strings or merge data structures. Complementary to the split block for data assembly.

sort

Sort arrays and objects by specified criteria. Essential for data organization and ranking operations.

batch

Group individual messages into batches for efficient processing. Improves performance for bulk operations.

Parser Blocks

csv

Parse CSV files and generate CSV output. Handles comma-separated value data processing and export.

html

Parse HTML content and extract structured data. Useful for web scraping and HTML document processing.

json

Parse JSON data and generate JSON output. Essential for working with structured data and API integration.

xml

Process XML documents and generate XML output. Handles structured document processing and data exchange.

yaml

Parse YAML configuration files and generate YAML output. Useful for configuration management and data serialization.

Storage Blocks

write file

Write data to files on the filesystem. Essential for persisting processed data and creating output files.

read file

Read data from files on the filesystem. Enables processing of stored data and configuration files.

watch

Monitor filesystem changes and trigger workflows when files are modified. Enables event-driven file processing.

Analysis Blocks

Large Language Model Blocks

Hallucination Detection

Detect and flag potential hallucinations in LLM outputs. Provides quality control for AI-generated content.

LLM Guard

Implement security and safety measures for Large Language Model interactions. Protects against harmful or inappropriate content.

create_conversation

Establish conversation contexts for AI interactions. Manages conversation state and history for chatbot-like applications.

fetch_history

Retrieve conversation history and previous interactions. Enables context-aware AI responses and conversation continuity.

LLM Query

Query Large Language Models for text generation, analysis, and reasoning. Core block for AI-powered text processing.

llm_query_v2

Enhanced version of LLM Query with improved capabilities and performance. Provides advanced LLM interaction features.

Pre-Process Document

Prepare documents for LLM processing by cleaning, formatting, and optimizing content. Essential preprocessing step for document workflows.

Pre-Process Document 2

Advanced document preprocessing with enhanced capabilities. Improved version of the document preprocessor.

Retrieve Matching Chunks

Find and retrieve relevant sections from large documents using similarity search. Core component for RAG (Retrieval Augmented Generation) systems.

Image Tools Blocks

viewer

Display data, results, and visualizations in the flow editor. Essential for monitoring workflow outputs and debugging.

image

Display and manipulate images within workflows. Handles image rendering and basic image operations.

Barcode Decoder

Decode barcodes and QR codes from images. Extracts encoded data for inventory and tracking applications.

Barcode Generator

Generate barcodes and QR codes. Creates encoded data representations for printing and digital display.

Agents Blocks

agent workgroup

Coordinate multiple AI agents to work together on complex tasks. Enables multi-agent workflows and collaboration patterns.

Multi Modal Blocks

Document Classifier

Automatically classify documents into predefined categories. Enables automated document sorting and routing.

Document Question Answering

Extract specific answers from documents based on questions. Enables intelligent document search and information extraction.

Document Understander

Advanced document comprehension and analysis. Provides deep understanding of document content and structure.

NLP Classifier

Classify text content using Natural Language Processing techniques. Enables topic classification and content categorization.

NLP Blocks

Entity Extractor

Extract named entities (people, places, organizations, dates) from text. Essential for information extraction and data structuring.

Computer Vision Blocks

Image Classifier

Classify images into predefined categories using machine learning models. Enables automated image sorting and content analysis.

Image Processor

Perform image manipulation, enhancement, and transformation operations. Handles image preprocessing and optimization.

Object Detector

Detect and locate objects within images. Provides bounding box coordinates and confidence scores for detected objects.

OCR

Optical Character Recognition to extract text from images and scanned documents. Essential for digitizing physical documents.

OCR Utils

Utility functions for OCR processing and text extraction. Provides additional tools for OCR workflows.

PDF Processor

Extract text, images, and metadata from PDF documents. Handles complex PDF processing and data extraction.

PDF Utils

Utility functions for PDF processing and manipulation. Provides additional tools for PDF workflows.

Signature Matcher

Detect and verify signatures in documents. Essential for document authentication and verification workflows.

Table Structure Recognition

Detect and extract table structures from documents and images. Converts tabular data into structured formats.

Tiff Utils

Process TIFF image files with specialized utilities. Handles high-quality image processing for document scanning workflows.

Template Matcher

Match documents against predefined templates to extract structured data. Enables form processing and data extraction.

LLM Guard Blocks

LLM Judge

Use Large Language Models for evaluation, scoring, and quality assessment. Enables AI-powered content evaluation.

General Blocks

verify_me

Human-in-the-loop validation system for AI-generated information. Enables human reviewers to validate AI results against source documents in an interactive viewer interface.

write_data

Store processed data to various destinations including databases, files, and external systems. Essential for persisting workflow results.

Visual Language Model Blocks

vlm_query

Query Vision Language Models for image analysis and understanding. Combines computer vision with natural language processing.

---

Usage Tips

• Combine Blocks: Most workflows require multiple blocks working together
• Start Simple: Begin with basic blocks before adding complex AI components
• Test Incrementally: Use debug blocks to verify each step of your workflow
• Read Documentation: Each block has detailed configuration options and examples

For detailed configuration information, click on any block name to view its complete documentation.