Advanced Workflows

Input → If (condition) → True Branch → Output A
                       → False Branch → Output B

Condition 1: {{word_count}} gt 100     AND
Condition 2: {{language}} eq "English"

Input → If (is_email?)
         → True → If (is_urgent?)
                    → True → Priority Handler
                    → False → Normal Handler
         → False → If (is_code?)
                     → True → Code Analyzer
                     → False → General AI

Input (list) → Loop Items → Process Each Item → Collect Results → Output

URL List → Loop Items (for each url)
             → Web Fetch ({{current_item}})
             → Summarize ({{fetch_result}})
           → Collect Summaries → Format Report

{{node_id.output}}           -- Full output of a node
{{node_id.output.field}}     -- Specific field from JSON output

{
  "target_language": "Spanish",
  "max_length": 500,
  "output_format": "markdown"
}

{% if word_count > 1000 %}
Summarize in 5 bullet points.
{% else %}
Provide a brief one-paragraph summary.
{% endif %}

{% for item in items %}
- {{ item.title }}: {{ item.description }}
{% endfor %}

Fetch API → Delay (5 seconds) → Fetch Next API

Input → [Parallel]
          → Translate to Spanish
          → Translate to French
          → Translate to German
        → [Merge] → Format All Translations

Main Workflow → Trigger Sub-Workflow ("summarize-and-translate")
             → Use Sub-Workflow Result
             → Continue Main Workflow

Node Settings:
  On Failure: retry
  Max Retries: 3
  Retry Delay: 2 seconds

Web Fetch → (success) → Process Content
          → (error) → Use Cached Content → Process Content

curl -X POST https://api.example.com/data \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query": "{{input_text}}"}'

Web Fetch (url: "https://api.example.com/news?q={{topic}}")
  → Parse JSON Response
  → Process Results

User Request → LLM ("Analyze this request and output JSON with: action, parameters")
             → Parse AI Output
             → Dynamic Action Node (uses AI-generated parameters)

{
  "action": "web_search_then_slack",
  "search_query": "today's top news",
  "channel": "#general",
  "format": "bullet_points"
}

Watch Files Changes (path: ~/Documents/notes/)
  → Read Changed File
  → Process Content
  → Update Knowledge Base

Trigger Schedule (every hour)
  → Gmail Search ("is:unread after:1h")
  → Loop Items (for each email)
      → Read Email Content
      → LLM ("Summarize in 2 sentences: {{email.body}}")
      → Translate (to: Spanish)
  → Collect Results
  → Format as HTML Report
  → Gmail Send (to: team@company.com, subject: "Hourly Email Digest")

Trigger Schedule (daily at 8 AM)
  → Loop Items (for each topic in ["AI", "Startups", "Tech"])
      → Web Search ("{{current_item}} news today")
      → Loop Items (for each result, max 5)
          → Web Fetch ({{result.url}})
          → LLM ("Extract key facts and insights")
      → Merge Results
  → Format as Daily Digest
  → Knowledge Base Ingest (collection: "Daily News")
  → IM Channel Send (Slack: #news-digest)

Input (code snippet or file path)
  → If (is_file_path?)
      → True → File System Read
      → False → Use Input Directly
  → [Parallel]
      → LLM ("Review for bugs and logic errors: {{code}}")
      → LLM ("Review for security vulnerabilities: {{code}}")
      → LLM ("Review for performance issues: {{code}}")
  → [Merge]
  → LLM ("Synthesize these reviews into a single report with severity ratings")
  → Output as Markdown

Input (folder path)
  → File System Glob ("{{folder}}/*.{png,jpg,jpeg}")
  → Loop Items (for each image)
      → [Parallel]
          → Image Compress (quality: 80%)
          → OCR (extract text)
      → If (has_text?)
          → True → Save OCR text to {{image_name}}.txt
          → False → Skip
  → Collect Results
  → Format Summary ("Processed {{count}} images, {{compressed_size}} saved")

IM Channel Listener (Slack: #support)
  → LLM ("Categorize this message: {{message}}
          Categories: billing, technical, feature_request, other
          Output JSON: {category, urgency, suggested_response}")
  → If (urgency eq "high")
      → True → IM Channel Send (Slack: #urgent-support, "URGENT: {{message}}")
  → If (category eq "billing")
      → True → LLM (use billing FAQ context)
      → False → If (category eq "technical")
          → True → LLM (use technical docs context + Knowledge Base)
          → False → LLM (general helpful response)
  → IM Channel Reply (thread: {{message.thread}}, content: {{response}})
  → Log to Knowledge Base (for future reference)