Items Reference

Use cases, dynamic items, collecting results, and best practices for the items: field.

Use Cases

Batch LLM Processing

# resources/batch-queries.yaml
actionId: batchQueries
items:
  - "What is machine learning?"
  - "Explain neural networks"
  - "What is deep learning?"

chat:
  prompt: "{{ get('current') }}"
  jsonResponse: true
  jsonResponseKeys:
    - answer

Data Enrichment

# resources/enrich-products.yaml
actionId: enrichProducts
requires: [fetchProducts]
items:
  - "get('fetchProducts')"

httpClient:
  method: GET
  url: "https://api.example.com/details/{{ get('current') }}"

Report Generation

# resources/generate-report.yaml
actionId: generateReport
items:
  - "executive_summary|Executive Summary"
  - "analysis|Market Analysis"
  - "recommendations|Recommendations"
  - "conclusion|Conclusion"

chat:
  prompt: |
    Generate the "{{ split(get('current'), '|')[1] }}" section of the report.
    Data: {{ get('reportData') }}

    {{ get('prev') ? 'Previous section: ' + get('prev') : '' }}
  jsonResponse: true
  jsonResponseKeys:
    - content
    - key_points

Multi-Language Translation

# resources/translate.yaml
actionId: translate
items:
  - "es|Spanish"
  - "fr|French"
  - "de|German"
  - "ja|Japanese"

chat:
  prompt: |
    Translate to {{ split(get('current'), '|')[1] }}:
    "{{ get('originalText') }}"
  jsonResponse: true
  jsonResponseKeys:
    - translation
    - language_code

Sequential Processing

# resources/chained-process.yaml
actionId: chainedProcess
items:
  - "1|gather_requirements"
  - "2|design_solution"
  - "3|implement"
  - "4|test"

chat:
  prompt: |
    Step {{ split(get('current'), '|')[0] }}: {{ split(get('current'), '|')[1] }}

    {{ get('prev') ? 'Previous step output: ' + get('prev') : 'Starting fresh.' }}

    Complete this step.

Image Batch Processing

# resources/analyze-images.yaml
actionId: analyzeImages
items:
  - "/uploads/image1.jpg"
  - "/uploads/image2.jpg"
  - "/uploads/image3.jpg"

chat:
  prompt: "Describe this image"
  files:
    - "{{ get('current') }}"

Collecting Results

Item results are collected into an array:

# Processing resource
actionId: processItems
items:
  - "Item 1"
  - "Item 2"

chat:
  prompt: "Process {{ get('current') }}"

---
# Response resource
actionId: response
requires: [processItems]
apiResponse:
  response:
    results: get('processItems')      # All results as array
    first: get('processItems')[0]     # First result
    count: len(get('processItems'))   # Result count

Dynamic Items

Items can come from expressions or previous resources. The items: field is always a YAML list; a single expression that returns an array is expanded into multiple iterations:

# From expression - evaluates to array, each element becomes one iteration
items:
  - "split(get('csv_data'), ',')"

# From previous resource output
items:
  - "get('fetchItems')"

# Filtered items
items:
  - "filter(get('allItems'), .status == 'active')"

Performance Considerations

• Items process sequentially - each waits for the previous to complete
• Set timeout to account for total processing time across all items
• For large datasets, Python batch processing is more efficient than items

Best Practices

• Use skip conditions to avoid processing items that don't qualify
• Use prev / next accessors to pass context between items
• For very large batches, use the python: resource with pandas

See Also

• Items - Core items concept and syntax
• Expressions - Dynamic item expressions
• Python Resource - Batch processing alternative