1Simple workflow:
2Trigger → Action → Action → Done
3 
4Advanced workflow:
5Trigger → Router → Multiple paths
6       → Error handling
7       → Iterations
8       → Sub-scenarios
9       → Data transformation
10       → Conditional logic

1Split flow into multiple paths:
2 
3                  ┌→ Path 1: If order
4Incoming  → Router┼→ Path 2: If question
5Message           └→ Path 3: Default
6 
7Each path:
8- Has filter condition
9- Executes independently
10- Can have different actions

1Filter conditions:
2- Contains: text contains "order"
3- Equals: status = "pending"
4- Greater than: amount > 100
5- Matches pattern: regex match
6- Is empty: field is blank
7- Exists: field has value

1AND: All must be true
2- status = "active"
3- AND amount > 50
4- AND type = "subscription"
5 
6OR: Any can be true
7- source = "WhatsApp"
8- OR source = "Telegram"
9- OR source = "SMS"

1Process arrays one item at a time:
2 
3Input: [item1, item2, item3]
4          ↓
5      Iterator
6          ↓
7    item1 → Process
8    item2 → Process
9    item3 → Process
10 
11Use for:
12- Process list of orders
13- Send multiple messages
14- Update multiple records

1Combine multiple items back:
2 
3    item1 → 
4    item2 → Aggregator → Combined result
5    item3 →
6 
7Types:
8- Array: Collect into array
9- Text: Join with separator
10- Numeric: Sum, average, count

1Scenario:
21. Get orders from API (returns array)
32. Iterator: For each order
43. Check inventory
54. Update status
65. Send notification
76. Aggregator: Collect results
87. Send summary report

1Common errors:
2- Connection failed
3- Invalid data
4- Rate limit
5- Timeout
6- Permission denied
7- Not found

1Stop execution on error:
21. Add Error Handler
32. Select "Break"
43. Execution stops
54. Error logged
65. No further actions

1Continue despite error:
21. Add Error Handler
32. Select "Resume"
43. Log error
54. Continue to next module
65. Workflow completes

1Save progress on error:
21. Add Error Handler
32. Select "Commit"
43. Data up to error point saved
54. Execution stops
65. Can resume later

1Undo changes on error:
21. Add Error Handler
32. Select "Rollback"
43. Undo committed changes
54. Clean state
65. Retry possible

1Handle specific errors:
2 
3Router (error path):
4If error code = 429 (rate limit)
5  → Wait 60 seconds
6  → Retry
7 
8If error code = 404 (not found)
9  → Create new record
10  → Continue
11 
12Default:
13  → Log error
14  → Send alert
15  → Stop

1Text:
2- lower(text): Lowercase
3- upper(text): Uppercase
4- trim(text): Remove whitespace
5- replace(text, old, new): Replace
6- substring(text, start, length): Extract
7 
8Numbers:
9- round(num, decimals)
10- floor(num), ceil(num)
11- formatNumber(num, format)
12 
13Dates:
14- formatDate(date, format)
15- addDays(date, days)
16- parseDate(text, format)
17 
18Arrays:
19- length(array)
20- first(array), last(array)
21- slice(array, start, end)
22- map(array, key)

1Parse JSON:
2- Use JSON module
3- parseJSON(text) function
4 
5Create JSON:
6- Use JSON > Create JSON
7- Build object from variables
8 
9Transform:
10- Extract specific fields
11- Restructure data
12- Filter arrays

1API returns:
2{
3  "data": {
4    "orders": [
5      {"id": 1, "total": 100, "status": "pending"},
6      {"id": 2, "total": 200, "status": "shipped"}
7    ]
8  }
9}
10 
11Transform:
121. Parse JSON
132. Access: data.orders
143. Filter: status = "pending"
154. Map: Extract only id and total
165. Result: [{"id": 1, "total": 100}]

1Benefits:
2- Reusable components
3- Cleaner main flow
4- Easier testing
5- Better organization

11. Create new scenario
22. Use webhook as trigger
33. Build your logic
44. Return result with webhook response
5 
6Main scenario:
71. Call sub-scenario via HTTP
82. Wait for response
93. Continue with result

1Sub-scenario: process_with_ai
2Input: {message, context}
3 
41. Webhook trigger
52. OpenAI: Generate response
63. Store to database
74. Webhook response: {reply, tokens_used}
8 
9Main scenarios call this sub-scenario
10when AI processing needed

1Create persistent storage:
21. Go to Data Stores
32. Create new
43. Define structure:
5   - Key (unique identifier)
6   - Fields (columns)

1Add/Update Record:
2- Add or replace by key
3- Upsert pattern
4 
5Search Records:
6- Filter by conditions
7- Get multiple results
8 
9Get Record:
10- Retrieve by key
11- Fast lookup
12 
13Delete Record:
14- Remove by key
15- Clean up data

1- Cache API responses
2- Track conversation state
3- Store user preferences
4- Rate limiting counters
5- Temporary data storage

1Handle transient failures:
2 
31. Try action
42. If fails:
5   a. Increment counter
6   b. If counter < 3:
7      - Wait (exponential backoff)
8      - Retry
9   c. Else:
10      - Log failure
11      - Alert

1Respect API limits:
2 
3Before API call:
41. Get counter from data store
52. If counter > limit:
6   - Wait until reset
73. Else:
8   - Increment counter
9   - Make call

1Prevent duplicate processing:
2 
31. Generate hash of incoming data
42. Check data store for hash
53. If exists:
6   - Skip (already processed)
74. Else:
8   - Store hash
9   - Process

1Interval:
2- Every X minutes
3- Every X hours
4 
5Time-based:
6- Every day at 9am
7- Every Monday
8- First of month
9 
10On-demand:
11- Webhook trigger
12- Manual run

1Example: Daily summary
2 
3Schedule: Every day at 6pm
41. Get today's conversations
52. Send to AI: "Summarize trends"
63. Generate report
74. Email to team

1Review past runs:
2- Input data
3- Output data
4- Errors
5- Duration
6- Module by module