MotteAF provides a drag-and-drop interface for designing complex multi-agent workflows with conditional logic, parallel processing, and dynamic routing.
Building blocks for agent workflows:
Advanced routing capabilities:
High-performance workflow execution with automatic scaling, error handling, and retry logic.
Flexible configuration system with environment variables, secrets management, and dynamic parameters.
| Node Type | Purpose | Inputs | Outputs |
|---|---|---|---|
| LLM Agent | Generate text responses | Prompt, context | Generated text |
| API Call | External service integration | URL, parameters | API response |
| Condition | Conditional branching | Data, condition | Boolean result |
| Transform | Data manipulation | Raw data | Processed data |
| Memory | Store/retrieve context | Query, data | Retrieved memories |
Input → Agent 1 → Transform → Agent 2 → Output
Linear workflow where each step depends on the previous one. Best for document processing, content generation pipelines.
Input → Split → [Agent A, Agent B, Agent C] → Merge → Output
Concurrent execution for independent tasks. Ideal for multi-perspective analysis, batch processing.
Input → Classifier → [Route A | Route B | Route C] → Output
Dynamic path selection based on input characteristics. Perfect for customer support routing, content categorization.
Input → Agent → Validator → [Continue | Retry] → Output
Iterative refinement with quality checks. Useful for content improvement, error correction workflows.
{
"name": "Customer Support Workflow",
"version": "1.0",
"nodes": [
{
"id": "classifier",
"type": "llm_agent",
"config": {
"model": "gpt-4",
"prompt": "Classify this customer inquiry: {{input}}",
"temperature": 0.1
}
},
{
"id": "technical_support",
"type": "llm_agent",
"config": {
"model": "gpt-4",
"prompt": "Provide technical support for: {{input}}",
"tools": ["knowledge_base", "ticket_system"]
}
}
],
"edges": [
{
"from": "classifier",
"to": "technical_support",
"condition": "output.category == 'technical'"
}
]
}POST /api/workflows/create
Content-Type: application/json
{
"name": "My Workflow",
"description": "Custom workflow description",
"definition": { ... workflow schema ... }
}Creates a new workflow from the provided definition.
POST /api/workflows/execute
Content-Type: application/json
{
"workflowId": "workflow_123",
"input": {
"message": "Customer inquiry text",
"metadata": { "priority": "high" }
}
}Executes a workflow with the provided input data.