Data Nodes

Data nodes handle input, output, and data manipulation in your workflows. Use them to accept files and URLs, produce outputs, and manage HTTP requests and database operations.

Input Nodes

Input Node

Basic input/constants for workflow initialization.

{
  inputType: "text",  // or "json"
  value: "Default value"
}

Files Node

Accept file uploads in your workflow.

Configuration

{
  inputName: "documents",
  required: true,
  allowedFileTypes: [
    "application/pdf",
    "text/plain",
    "text/markdown",
    "application/vnd.openxmlformats-officedocument.wordprocessingml.document"
  ],
  maxFiles: 10,
  maxFileSizeMb: 25,
  chunkSize: 1000,
  chunkOverlap: 200,
  outputVariable: "uploaded_files",
  label: "Upload Documents",
  description: "Upload PDF, TXT, MD, or DOCX files"
}

Supported File Types

Category	Extensions	MIME Types
Documents	PDF, DOCX, DOC, TXT, MD, HTML	application/pdf, etc.
Data	JSON, CSV, XLSX	application/json, text/csv
Code	JS, TS, PY, etc.	text/javascript, etc.
Images	PNG, JPG, GIF, WEBP	image/*

Output

{
    files: [
        {
            name: "report.pdf",
            size: 1024000,
            mimeType: "application/pdf",
            content: "Extracted text content...",
            chunks: [
                { text: "Chunk 1...", index: 0 },
                { text: "Chunk 2...", index: 1 }
            ]
        }
    ];
}

Chunking

Files are automatically chunked for RAG applications:

• chunkSize: Characters per chunk (default: 1000)
• chunkOverlap: Overlap between chunks (default: 200)

---

URL Node

Accept and process URLs.

{
  inputName: "urls",
  scrapingMode: "markdown",  // html, text, markdown
  scrapeSubpages: false,
  timeout: 30,
  followRedirects: true,
  includeMetadata: true,
  chunkingAlgorithm: "semantic",
  chunkSize: 1000,
  chunkOverlap: 200,
  advancedExtraction: false,
  ocrEnabled: false,
  outputVariable: "url_content",
  label: "Reference URLs",
  description: "Enter URLs to include as context"
}

Scraping Modes

Mode	Output
html	Raw HTML content
text	Plain text only
markdown	Structured markdown

Output

{
  urls: [
    {
      url: "https://example.com/article",
      title: "Article Title",
      content: "Extracted content...",
      metadata: {
        description: "...",
        author: "...",
        publishedDate: "..."
      },
      chunks: [...]
    }
  ]
}

---

Web Search Node

Search the web and return results.

{
  query: "{{user.search_term}}",
  maxResults: 10,
  searchType: "general",  // general, news, images
  outputVariable: "search_results"
}

Output

{
    results: [
        {
            title: "Result Title",
            url: "https://...",
            snippet: "Preview text...",
            publishedDate: "2024-01-15"
        }
    ];
}

---

Web Browse Node

Fetch and read web page content.

{
  url: "{{input.url}}",
  extractText: true,
  maxLength: 50000,
  outputVariable: "page_content"
}

---

PDF Extract Node

Extract content from PDF files.

{
  path: "{{files_node.output.files[0].url}}",
  extractText: true,
  extractMetadata: true,
  pageStart: 1,
  pageEnd: null,  // All pages
  specificPages: null,  // Or [1, 3, 5]
  outputFormat: "text",
  password: null,  // For encrypted PDFs
  outputVariable: "pdf_content"
}

---

File Download Node

Download files from URLs.

{
  url: "{{api_response.file_url}}",
  filename: "downloaded_file.pdf",
  maxSize: 52428800,  // 50MB
  timeout: 60,
  followRedirects: true,
  allowedContentTypes: ["application/pdf", "image/*"],
  outputVariable: "downloaded_file"
}

---

File Read Node

Read files from storage.

{
  path: "{{file.path}}",
  encoding: "utf-8",  // utf-8, base64, binary
  maxSize: 10485760,  // 10MB
  outputVariable: "file_content"
}

---

Output Nodes

Output Node

Collect and return workflow outputs.

{
  outputName: "result",
  value: "{{llm_node.output.response}}",
  format: "json",  // json, string, number, boolean
  description: "The main workflow output"
}

Multiple output nodes create a combined output object:

{
  result: "...",
  summary: "...",
  metadata: {...}
}

---

Template Output Node

Render templates with variable interpolation.

{
  outputName: "report",
  template: `
# Report for {{company.name}}

Generated on: {{current_date}}

## Summary
{{analysis.summary}}

## Key Findings
{{#each findings}}
- **{{this.title}}**: {{this.description}}
{{/each}}

## Recommendations
{{recommendations}}
  `,
  outputFormat: "markdown",  // markdown, html
  description: "Formatted report"
}

Template Syntax

Syntax	Description
{{variable}}	Insert variable value
{{#each items}}...{{/each}}	Loop over array
{{#if condition}}...{{/if}}	Conditional content
{{this}}	Current item in loop
{{@index}}	Current loop index

---

Action Node

Perform external actions (send, create, update, delete).

{
  provider: "slack",
  operation: "send_message",
  connectionId: "conn_abc123",
  parameters: {
    channel: "#general",
    message: "{{notification.text}}"
  },
  outputVariable: "action_result"
}

Used for:

• Sending notifications
• Creating records
• Updating external systems
• Triggering integrations

---

Chart Generation Node

Create charts and visualizations.

{
  chartType: "bar",
  dataSource: "{{analytics.monthly_data}}",
  dataLabels: {
    x: "month",
    y: "revenue"
  },
  title: "Monthly Revenue",
  subtitle: "2024",
  xAxisLabel: "Month",
  yAxisLabel: "Revenue ($)",
  width: 800,
  height: 400,
  theme: "light",
  legend: true,
  showGrid: true,
  showValues: true,
  filename: "revenue_chart.png",
  outputVariable: "chart_output"
}

Chart Types

Type	Description
bar	Vertical bar chart
horizontal_bar	Horizontal bar chart
line	Line chart
area	Area chart
pie	Pie chart
donut	Donut chart
scatter	Scatter plot
histogram	Histogram
heatmap	Heat map

---

Spreadsheet Generation Node

Create spreadsheets from data.

{
  format: "xlsx",  // xlsx, csv
  filename: "export.xlsx",
  dataSource: "{{query_results}}",
  sheetName: "Data",
  headerBold: true,
  headerBackgroundColor: "#4472C4",
  headerFontColor: "#FFFFFF",
  alternateRows: true,
  freezeHeader: true,
  outputVariable: "spreadsheet_output"
}

---

PDF Generation Node

Create PDF documents.

{
  content: "{{template_output.rendered}}",
  format: "markdown",  // markdown, html
  filename: "document.pdf",
  pageSize: "A4",
  orientation: "portrait",
  margins: { top: 20, right: 20, bottom: 20, left: 20 },
  headerText: "Company Name",
  footerText: "Confidential",
  includePageNumbers: true,
  outputVariable: "pdf_output"
}

---

Screenshot Capture Node

Capture screenshots of web pages.

{
  url: "{{input.url}}",
  fullPage: true,
  width: 1920,
  height: 1080,
  deviceScale: 2,
  format: "png",  // png, jpeg, webp
  quality: 90,
  delay: 1000,
  selector: null,  // Specific element
  darkMode: false,
  timeout: 30000,
  filename: "screenshot.png",
  outputVariable: "screenshot_output"
}

---

File Write Node

Write content to files.

{
  path: "output/report.txt",
  content: "{{generated_content}}",
  encoding: "utf-8",
  createDirectories: true,
  overwrite: true,
  outputVariable: "file_write_result"
}

---

HTTP Node

Make HTTP requests to external APIs.

Configuration

{
  method: "POST",
  url: "https://api.example.com/endpoint",
  headers: {
    "Content-Type": "application/json",
    "X-Custom-Header": "{{custom_value}}"
  },
  queryParams: {
    "filter": "active",
    "limit": "100"
  },
  authType: "bearer",
  authCredentials: {
    token: "{{env.API_TOKEN}}"
  },
  bodyType: "json",
  body: {
    "name": "{{user.name}}",
    "email": "{{user.email}}"
  },
  timeout: 30000,
  retryCount: 3,
  outputVariable: "api_response"
}

Methods

Method	Use Case
GET	Retrieve data
POST	Create resources
PUT	Replace resources
PATCH	Partial updates
DELETE	Remove resources

Authentication Types

Type	Configuration
none	No authentication
basic	{ username, password }
bearer	{ token }
apiKey	{ key, value, location }

Body Types

Type	Content-Type
json	application/json
form	application/x-www-form-urlencoded
raw	Custom (set in headers)

Output

{
  status: 200,
  statusText: "OK",
  headers: {...},
  body: {...},
  responseTime: 245
}

---

Database Node

Query and manipulate databases.

Supported Databases

Type	Description
postgresql	PostgreSQL
mysql	MySQL / MariaDB
mongodb	MongoDB

Operations

Operation	Description
query	Execute SELECT query
insert	Insert records
update	Update records
delete	Delete records

Configuration

{
  databaseType: "postgresql",
  connectionId: "conn_database",
  operation: "query",
  query: "SELECT * FROM users WHERE status = $1 AND created_at > $2",
  parameters: ["active", "2024-01-01"],
  outputVariable: "query_results"
}

Insert Example

{
  operation: "insert",
  query: "INSERT INTO logs (event, data, timestamp) VALUES ($1, $2, NOW())",
  parameters: ["user_action", "{{event_data}}"]
}

MongoDB Example

{
  databaseType: "mongodb",
  operation: "query",
  query: {
    collection: "users",
    filter: { status: "active" },
    projection: { name: 1, email: 1 },
    sort: { createdAt: -1 },
    limit: 100
  }
}

---

Best Practices

Input Handling

• Validate file types and sizes
• Set appropriate chunk sizes for RAG
• Use descriptive labels and descriptions
• Handle empty inputs gracefully

Output Formatting

• Use template outputs for structured documents
• Choose appropriate chart types for data
• Compress large outputs when possible
• Include metadata for traceability

HTTP Requests

• Always set timeouts
• Use retry for transient failures
• Secure credentials in environment variables
• Validate responses before use

Database Operations

• Use parameterized queries (prevent SQL injection)
• Limit result sets for large tables
• Handle connection errors
• Log operations for debugging