Using AI Blocks

Detailed explanation of the functions and configuration methods of each AI block in nanocraft

We are continuously adding new AI blocks to enhance your workflow automation experience. More blocks and features will be available soon.

Block Categories

Input Blocks

Data acquisition and loading

7 types of blocks

Process Blocks

AI data processing and analysis

7 types of blocks

Output Blocks

Send to external services

5 types of blocks

Input Block Details

Detailed explanation of blocks for data acquisition and loading

Gmail Receive

Input Block

Retrieves emails from Gmail and passes information such as subject, body, and sender to subsequent blocks.

Configuration items:

  • • Email types (unread, read, sent, etc.)
  • • Number of items to retrieve (max 100)
  • • Search query (filter by specific conditions)

Audio Input & Acoustic Input

Input Block

Processes audio files and recorded acoustic data using AI. Audio input provides speech-to-text conversion, while acoustic input analyzes music, environmental sounds, mechanical sounds, etc.

Features:

  • • Speech recognition (Speech-to-Text)
  • • Acoustic analysis (music, environmental sounds, mechanical sounds, etc.)
  • • Input methods (microphone recording / file selection)

File Reading

Input Block

Reads PDF, Excel, text files, etc., extracts content and converts it into a format that can be processed by subsequent blocks.

Supported formats:

  • • PDF (text extraction)
  • • Excel/CSV (tabular data)
  • • Word documents
  • • Text files

Image Input

Input Block

Analyzes image files using AI to provide content descriptions, text recognition (OCR), object detection, etc.

Features:

  • • Automatic image content description generation
  • • Text recognition (OCR)
  • • Object and person detection
  • • Color analysis

API Call

Input Block

Calls external service HTTP APIs to retrieve data. Supports various API formats such as REST API and GraphQL.

Configuration items:

  • • API endpoint URL
  • • HTTP method (GET, POST, PUT, DELETE)
  • • Header settings (authentication information, etc.)
  • • Request body (JSON, etc.)

Google Drive Retrieve

Input Block

Retrieves files from Google Drive and extracts their content. Supports multiple file formats including Google Docs, Spreadsheets, PDFs, etc.

Supported files:

  • • Google Docs, Spreadsheets, Slides
  • • PDF, Word, Excel, PowerPoint
  • • Text files, image files

Web Scraping

Input Block

Retrieves and analyzes content from web pages at specified URLs. Can collect various web content from news sites, blogs, e-commerce sites, etc.

Retrieved data:

  • • Page title, body text
  • • Metadata (author, publication date, etc.)
  • • Image URLs, link information
  • • Structured data (JSON-LD, etc.)

Process Block Details

Detailed explanation of blocks for data processing and analysis using AI

Text Summarization

Process Block

AI automatically summarizes long texts and multiple email bodies to extract important points.

Features:

  • • Automatic text summarization with optimal length
  • • Support for various text formats (emails, documents, web content)
  • • Japanese language optimized processing
  • • Maintains important context and key points

Sentiment Analysis

Process Block

Analyzes text emotions and determines positive/negative/neutral sentiment along with confidence measurement.

Output results:

  • • Sentiment classification (positive/negative/neutral)
  • • Confidence score (0.0-1.0)
  • • Emotion types (joy, anger, sadness, etc.)
  • • Analysis explanation

Content Classification

Process Block

Automatically categorizes text content and classifies it into genres such as business, technology, personal, etc.

Classification examples:

  • • Business (sales, marketing, etc.)
  • • Technology (development, infrastructure, etc.)
  • • Support (inquiries, issues, etc.)
  • • Others (personal, etc.)

Data Extraction

Process Block

Automatically extracts specific information (dates, names, amounts, contacts, etc.) from text and outputs it as structured data.

Extraction targets:

  • • Dates and times
  • • Person names and company names
  • • Amounts and numbers
  • • Email addresses and phone numbers

Conditional Branching

Process Block

Branches processing based on data content. Different processing flows can be executed based on conditions such as sentiment analysis results, amounts, categories, etc.

Condition setting examples:

  • • Emergency response if sentiment is negative
  • • Supervisor approval if amount is 50,000 yen or more
  • • Route to development team if category is 'technical'
  • • Immediate notification if priority is high

Usage Tips

By drawing multiple output lines from the conditional branching block, you can set different processes for each condition. Default branching can also be configured to handle unexpected data.

Custom Prompt

Process Block

Allows you to freely write AI prompts and execute custom AI processing. Enables special processing and advanced customization that standard processing blocks cannot handle.

Usage examples:

  • • Text analysis with custom business rules
  • • Classification using industry-specific terminology
  • • Complex workflows combining multiple processes
  • • Data formatting with specific output formats

Loop

Process Block

Repeatedly executes processing according to specified conditions or number of times. Ideal for large-scale data processing and automation of periodic tasks.

Loop types:

  • • while loop: Repeats while condition is true
  • • for loop: Executes for specified number of times or for each array element
  • • Scheduled execution: Auto-executes at time intervals or schedules
  • • Infinite loop prevention: Maximum execution count limit feature

Usage examples

  • • Process each email individually from email list
  • • Generate reports daily at fixed times
  • • Poll API until conditions are met

Output Block Details

Detailed explanation of blocks for sending processing results to external services

Slack Send

Output Block

Automatically sends processing results to Slack channels. Can instantly share summaries, analysis reports, alerts, etc. with teams.

Features:

  • • OAuth integration with Slack workspace
  • • Channel selection from available channels
  • • Automatic message formatting from workflow results
  • • Real-time message delivery

Google Sheets Save

Output Block

Automatically saves extracted data and analysis results to Google Sheets spreadsheets. Ideal for data accumulation and later analysis.

Functions:

  • • OAuth integration with Google Sheets
  • • Create new spreadsheet automatically
  • • Append data to existing spreadsheets
  • • CSV format data export

Notion Create

Output Block

Notion integration for creating pages in databases. Requires API token and database ID configuration.

機能：

  • • Basic page creation in Notion databases
  • • Direct integration with Notion API
  • • API token configuration required
  • • Database ID specification needed

PDF Output

Output Block

Outputs analysis results and reports as formatted PDF files. Suitable for printing, archiving, and creating formal reports.

Layout:

  • • Automatic header and footer generation
  • • Table of contents and page numbering
  • • Data table display and formatting
  • • Japanese font support and styling

Google Drive Save

Output Block

Automatically saves processing results and generated files to Google Drive. Can save in various file formats including text files, PDFs, spreadsheets, etc.

機能：

  • • Multiple file format support (TXT, JSON, CSV, MD, HTML, PDF)
  • • Folder specification with automatic folder creation
  • • PDF generation and upload integration
  • • OAuth-based Google Drive authentication

Best Practices

Tips for effective workflow creation

Recommended Usage

  • Start with small workflows and gradually increase complexity
  • Always test each block's configuration before saving
  • Place error handling blocks in appropriate locations
  • Give workflows clear names and descriptions

Things to Avoid

  • Running blocks with incomplete configuration
  • Creating connections that cause infinite loops
  • Attempting to process large amounts of data at once
  • Forgetting to configure authentication for services that require it

Next Steps

Once you understand how to use AI blocks, let's create an actual workflow

Create Workflow Troubleshooting