Agents & Connectors

Overview

Agents

Processing components that perform specific actions in document generation workflows. Agents can be:

Built-in agents: Pre-configured agents for common tasks (search, analysis, formatting)
Custom agents: User-defined agents with specific instructions and outputs

Connectors

Specialized tools that give agents access to data-specific tools and functions. Connectors enable agents to:

Interpret different document types within the generation pipeline
Access external systems and databases
Apply specific rules and functions to data processing

Important: Connectors operate after document ingestion and help the system understand various data types available to agents.

Connector Management

Create Connector

Endpoint

POST /connectors

Headers

Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

Request Body

{
  "name": "LIMSConnector",
  "sourceType": "database",
  "description": "Laboratory Information Management System connector",
  "config": {
    "auth_type": "oauth",
    "client_id": "your_client_id",
    "client_secret": "your_client_secret",
    "token_url": "https://lims.company.com/oauth/token",
    "api_endpoint": "https://lims.company.com/api/v1"
  },
  "functions": [
    "extract_lab_results",
    "map_specimen_data",
    "validate_test_parameters"
  ],
  "rules": [
    "Only process results with status 'completed'",
    "Map sample IDs to patient identifiers",
    "Validate date formats as ISO 8601"
  ]
}

Request Parameters

Field	Type	Required	Description
`name`	String	Yes	Human-readable connector name
`sourceType`	String	Yes	Type of data source: “database” or “document”
`description`	String	No	Connector description and purpose
`config`	Object	No	Configuration including authentication credentials
`functions`	Array[String]	No	List of functions the connector provides to agents
`rules`	Array[String]	No	Processing rules and constraints

Supported Source Types

database: Connects to external databases and data systems
document: Processes specific document formats and structures

Authentication Types in Config

// OAuth 2.0
{
  "auth_type": "oauth",
  "client_id": "your_client_id",
  "client_secret": "your_client_secret",
  "token_url": "https://api.example.com/oauth/token",
  "scope": "read:data"
}

// API Key
{
  "auth_type": "api_key",
  "api_key": "your_api_key",
  "key_header": "X-API-Key"
}

// Basic Authentication
{
  "auth_type": "basic",
  "username": "your_username",
  "password": "your_password"
}

Example Request

curl -X POST "https://api.artosai.com/connectors" \
  -H "Authorization: Bearer your-access-token" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "ClinicalTrialConnector",
    "sourceType": "database",
    "description": "Connects to clinical trial management system",
    "config": {
      "auth_type": "api_key",
      "api_key": "ctms_12345",
      "api_endpoint": "https://ctms.company.com/api"
    },
    "functions": [
      "extract_patient_data",
      "map_study_protocols",
      "validate_eligibility_criteria"
    ],
    "rules": [
      "Only access patients with signed consent",
      "Anonymize all patient identifiers",
      "Validate protocol compliance"
    ]
  }'

Success Response (201 Created)

{
  "id": "conn-abc123def456",
  "name": "ClinicalTrialConnector",
  "sourceType": "database",
  "status": "created",
  "created_at": "2024-01-15T10:30:00Z",
  "created_by": "user-xyz789"
}

List All Connectors

Endpoint

GET /connectors

Headers

Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

Query Parameters

Parameter	Type	Required	Description
`limit`	Integer	No	Results per page (default: 20, max: 100)
`offset`	Integer	No	Number of results to skip (default: 0)
`status`	String	No	Filter by status

Example Request

curl -X GET "https://api.artosai.com/connectors?sourceType=database" \
  -H "Authorization: Bearer your-access-token"

Success Response (200 OK)

[
  {
    "id": "conn-abc123def456",
    "name": "ClinicalTrialConnector",
    "sourceType": "database",
    "description": "Connects to clinical trial management system",
    "status": "active",
    "created_at": "2024-01-15T10:30:00Z",
    "created_by": "user-xyz789",
    "last_used": "2024-01-15T14:22:00Z",
    "usage_count": 15,
    "functions": [
      "extract_patient_data",
      "map_study_protocols",
      "validate_eligibility_criteria"
    ],
    "rules": [
      "Only access patients with signed consent",
      "Anonymize all patient identifiers",
      "Validate protocol compliance"
    ]
  },
  {
    "id": "conn-preset-fda-001",
    "name": "FDA Regulatory Connector",
    "sourceType": "document",
    "description": "Processes FDA regulatory documents and submissions",
    "status": "preset",
    "created_at": "2024-01-10T09:00:00Z",
    "created_by": "system",
    "last_used": "2024-01-15T16:45:00Z",
    "usage_count": 127,
    "functions": [
      "parse_regulatory_sections",
      "extract_submission_data",
      "validate_compliance_requirements"
    ],
    "rules": [
      "Follow FDA formatting guidelines",
      "Maintain regulatory traceability",
      "Ensure section completeness"
    ]
  }
]

Get Specific Connector

Endpoint

GET /connectors/{connectorId}

Path Parameters

Parameter	Type	Required	Description
`id`	String	Yes	Connector ID

Example Request

curl -X GET "https://api.artosai.com/connectors/conn-abc123def456" \
  -H "Authorization: Bearer your-access-token"

Success Response (200 OK)

{
  "id": "conn-abc123def456",
  "name": "ClinicalTrialConnector",
  "sourceType": "database",
  "description": "Connects to clinical trial management system",
  "status": "active",
  "config": {
    "auth_type": "api_key",
    "api_endpoint": "https://ctms.company.com/api",
    "connection_timeout": 30
  },
  "functions": [
    "extract_patient_data",
    "map_study_protocols", 
    "validate_eligibility_criteria"
  ],
  "rules": [
    "Only access patients with signed consent",
    "Anonymize all patient identifiers",
    "Validate protocol compliance"
  ],
  "created_at": "2024-01-15T10:30:00Z",
  "created_by": "user-xyz789",
  "last_updated": "2024-01-15T12:15:00Z",
  "last_used": "2024-01-15T14:22:00Z",
  "usage_count": 15,
  "connected_pipelines": 3
}

Update Connector

Endpoint

PUT /connectors/{connectorId}

Path Parameters

Parameter	Type	Required	Description
`id`	String	Yes	Connector ID

Request Body Same format as create connector request. Example Request

curl -X PUT "https://api.artosai.com/connectors/conn-abc123def456" \
  -H "Authorization: Bearer your-access-token" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Enhanced ClinicalTrialConnector",
    "sourceType": "database",
    "description": "Enhanced connector with additional validation rules",
    "config": {
      "auth_type": "oauth",
      "client_id": "updated_client_id",
      "client_secret": "updated_secret",
      "token_url": "https://ctms.company.com/oauth/token"
    },
    "functions": [
      "extract_patient_data",
      "map_study_protocols",
      "validate_eligibility_criteria",
      "generate_safety_reports"
    ],
    "rules": [
      "Only access patients with signed consent",
      "Anonymize all patient identifiers", 
      "Validate protocol compliance",
      "Generate audit trail for all actions"
    ]
  }'

Success Response (200 OK)

{
  "id": "conn-abc123def456",
  "name": "Enhanced ClinicalTrialConnector",
  "status": "updated",
  "updated_at": "2024-01-15T15:30:00Z",
  "version": 2
}

Delete Connector

Endpoint

DELETE /connectors/{connectorId}

Path Parameters

Parameter	Type	Required	Description
`id`	String	Yes	Connector ID

⚠️ Warning Deleting a connector will impact all pipelines that use this connector. Ensure no active pipelines depend on this connector before deletion. Example Request

curl -X DELETE https://api.artosai.com/connectors/conn_abc123def456 \
  -H "Authorization: Bearer YOUR_API_KEY"

Success Response (200 OK)

{
  "success": true,
  "message": "Connector deleted successfully"
}

Test Connector

Endpoint

POST /connectors/{connectorId}/test

Path Parameters

Parameter	Type	Required	Description
`id`	String	Yes	Connector ID

Example Request

curl -X POST https://api.artosai.com/connectors/conn_abc123def456/test \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "testData": {
      "data": [
        {"id": 1, "name": "Test Item", "value": 100.5},
        {"id": 2, "name": "Test Item 2", "value": 200.3}
      ],
      "format": "json"
    },
    "expectedOutput": {
      "processedRows": 2,
      "validationPassed": true
    }
  }'

Success Response (200 OK)

{
  "testId": "test_xyz789",
  "status": "passed",
  "results": {
    "processedRows": 2,
    "validationPassed": true,
    "processingTime": 0.5,
    "accuracy": 1.0
  },
  "details": {
    "agentSelection": ["data_validator", "table_processor"],
    "executionPath": "optimal",
    "performanceMetrics": {
      "latency": "0.5s",
      "throughput": "4 rows/s"
    }
  }
}
```\n\n## Agent Framework

### List Available Agents

**Endpoint**

GET /agents


**Headers**

Authorization: Bearer YOUR_API_KEY Content-Type: application/json

**Query Parameters**

| Parameter    | Type   | Required | Description                                             |
| ------------ | ------ | -------- | ------------------------------------------------------- |
| `type`       | String | No       | Filter by agent type: "builtin" or "custom"             |
| `category`   | String | No       | Filter by category (search, analysis, formatting, etc.) |
| `created_by` | String | No       | Filter by creator (for custom agents)                   |

**Example Request**

```bash
curl -X GET "https://api.artosai.com/agents?type=builtin" \
  -H "Authorization: Bearer your-access-token"

Success Response (200 OK)

[
  {
    "id": "agent-search-001",
    "name": "Document Search Agent",
    "type": "builtin",
    "category": "search",
    "description": "Searches and retrieves relevant passages from ingested documents",
    "capabilities": [
      "semantic_search",
      "keyword_matching",
      "relevance_ranking"
    ],
    "output_formats": ["text", "json"],
    "version": "1.2.0",
    "created_by": "system"
  },
  {
    "id": "agent-mapping-002",
    "name": "Data Mapping Agent",
    "type": "builtin",
    "category": "analysis",
    "description": "Maps data tables to document sections and structures",
    "capabilities": [
      "table_extraction",
      "section_mapping",
      "data_validation"
    ],
    "output_formats": ["html", "json"],
    "version": "1.1.0",
    "created_by": "system"
  },
  {
    "id": "agent-qc-003",
    "name": "Quality Control Agent",
    "type": "builtin",
    "category": "validation",
    "description": "Performs quality checks on generated content",
    "capabilities": [
      "content_validation",
      "consistency_checking",
      "compliance_verification"
    ],
    "output_formats": ["text", "html"],
    "version": "1.0.0",
    "created_by": "system"
  },
  {
    "id": "agent-custom-001",
    "name": "CustomAgent",
    "type": "builtin",
    "category": "custom",
    "description": "Configurable agent that executes user-defined instructions",
    "capabilities": [
      "custom_instructions",
      "flexible_processing",
      "user_defined_logic"
    ],
    "output_formats": ["text", "html", "markdown"],
    "version": "2.0.0",
    "created_by": "system"
  },
  {
    "id": "agent-user-custom-123",
    "name": "Regulatory Parser",
    "type": "custom",
    "category": "analysis",
    "description": "Custom agent for parsing regulatory documents",
    "capabilities": [
      "regulatory_parsing",
      "compliance_extraction"
    ],
    "output_formats": ["html"],
    "version": "1.0.0",
    "created_by": "user-xyz789",
    "created_at": "2024-01-15T11:00:00Z"
  }
]

Get Specific Agent

Endpoint

GET /agents/{agentId}

Path Parameters

Parameter	Type	Required	Description
`agentId`	String	Yes	Agent ID

Example Response

{
  "id": "agent-search-001",
  "name": "Document Search Agent",
  "type": "builtin",
  "category": "search",
  "description": "Searches and retrieves relevant passages from ingested documents",
  "capabilities": [
    "semantic_search",
    "keyword_matching",
    "relevance_ranking"
  ],
  "output_formats": ["text", "json"],
  "version": "1.2.0",
  "created_by": "system",
  "usage_count": 1247,
  "last_used": "2024-01-15T14:30:00Z",
  "compatible_connectors": [
    "conn-preset-fda-001",
    "conn-abc123def456"
  ]
}

Create Custom Agent

Endpoint

POST /agents

Headers

Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

Request Body

{
  "name": "RegDocParser",
  "description": "Custom agent for parsing regulatory documents and extracting key information",
  "category": "analysis",
  "instructions": "Parse regulatory documents to extract approval dates, indication statements, and safety warnings. Format output as structured HTML with clear section headers.",
  "output_format": "html"
}

Request Parameters

Field	Type	Required	Description
`name`	String	Yes	Human-readable agent name
`description`	String	Yes	Agent description and purpose
`category`	String	No	Agent category (analysis, formatting, validation, etc.)
`instructions`	String	Yes	Detailed instructions for agent behavior
`output_format`	String	Yes	Output format: “text”, “html”, or “markdown”

Example Request

curl -X POST "https://api.artosai.com/agents" \
  -H "Authorization: Bearer your-access-token" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Safety Report Analyzer",
    "description": "Analyzes clinical trial safety data and generates summary reports",
    "category": "analysis",
    "instructions": "Review safety data from clinical trials and generate a comprehensive summary including adverse events, serious adverse events, and safety signal analysis. Present findings in structured HTML format with tables for quantitative data.",
    "output_format": "html"
  }'

Success Response (201 Created)

{
  "id": "agent-user-custom-456",
  "name": "Safety Report Analyzer",
  "type": "custom",
  "status": "created",
  "created_at": "2024-01-15T12:00:00Z",
  "created_by": "user-xyz789"
}

Update Agent

Endpoint

PUT /agents/{agentId}

Path Parameters

Parameter	Type	Required	Description
`agentId`	String	Yes	Agent ID

Request Body Same format as create agent request. Note: Only custom agents created by the user can be updated. Built-in agents cannot be modified. Example Request

curl -X PUT "https://api.artosai.com/agents/agent-user-custom-456" \
  -H "Authorization: Bearer your-access-token" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Enhanced Safety Report Analyzer",
    "description": "Enhanced analyzer with additional statistical analysis capabilities",
    "category": "analysis",
    "instructions": "Review safety data from clinical trials and generate comprehensive summaries including adverse events, serious adverse events, safety signal analysis, and statistical significance testing. Include comparative analysis against historical controls when available.",
    "output_format": "html"
  }'

Success Response (200 OK)

{
  "id": "agent-user-custom-456",
  "name": "Enhanced Safety Report Analyzer",
  "status": "updated",
  "updated_at": "2024-01-15T13:00:00Z",
  "version": 2
}

Delete Agent

Endpoint

DELETE /agents/{agentId}

Path Parameters

Parameter	Type	Required	Description
`agentId`	String	Yes	Agent ID

Note: Only custom agents created by the user can be deleted. Built-in agents cannot be removed. Example Request

curl -X DELETE "https://api.artosai.com/agents/agent-user-custom-456" \
  -H "Authorization: Bearer your-access-token"

Success Response (200 OK)

{
  "message": "Agent successfully deleted",
  "agentId": "agent-user-custom-456",
  "deleted_at": "2024-01-15T16:30:00Z",
  "affected_pipelines": [
    "pipeline-789"
  ]
}

Call Agent Directly

Execute an agent outside of a pipeline for testing or standalone processing. Endpoint

POST /agents/{agentId}/call

Path Parameters

Parameter	Type	Required	Description
`agentId`	String	Yes	Agent ID

Headers

Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

Request Body

{
  "documentUrls": [
    "s3://your-bucket/ingested/doc-123/clinical-protocol.docx",
    "s3://your-bucket/ingested/doc-456/safety-data.csv"
  ],
  "instructions": "Focus on extraction of primary endpoints and safety measures"
}

Request Parameters

Field	Type	Required	Description
`documentUrls`	Array[String]	Yes	S3 URLs of ingested documents to process
`instructions`	String	No	Additional instructions for this specific call

Example Request

curl -X POST "https://api.artosai.com/agents/agent-search-001/call" \
  -H "Authorization: Bearer your-access-token" \
  -H "Content-Type: application/json" \
  -d '{
    "documentUrls": [
      "s3://your-bucket/ingested/doc-123/protocol.docx"
    ],
    "instructions": "Extract all mentions of primary and secondary endpoints"
  }'

Success Response (200 OK)

{
  "result": {
    "output": "Primary Endpoint: The primary endpoint is overall survival (OS) defined as time from randomization to death from any cause. Secondary Endpoints: 1) Progression-free survival (PFS), 2) Objective response rate (ORR), 3) Duration of response (DOR), 4) Quality of life measures using EORTC QLQ-C30.",
    "format": "text",
    "processing_time_seconds": 12.5,
    "documents_processed": 1,
    "connectors_used": [
      "conn-preset-fda-001"
    ]
  },
  "agent_id": "agent-search-001",
  "execution_id": "exec-abc123def456",
  "timestamp": "2024-01-15T14:45:00Z"
}

CustomAgent Usage

The CustomAgent is a special built-in agent that allows users to define their own processing logic through instructions.

Using CustomAgent in Pipelines

When creating a pipeline with CustomAgent:

{
  "name": "Custom Processing Pipeline",
  "agentIds": ["agent-search-001", "agent-custom-001", "agent-qc-003"],
  "order": [1, 2, 3],
  "customAgentInstructions": {
    "agent-custom-001": {
      "instructions": "Take the search results and format them into a regulatory submission format following FDA guidelines. Include all required sections and ensure proper citations.",
      "output_format": "html"
    }
  },
  "documents": ["doc-123", "doc-456"],
  "id": "conn-regulatory-001",
  "outputFileName": "regulatory-submission.docx"
}

CustomAgent Direct Call

curl -X POST "https://api.artosai.com/agents/agent-custom-001/call" \
  -H "Authorization: Bearer your-access-token" \
  -H "Content-Type: application/json" \
  -d '{
    "documentUrls": ["s3://bucket/doc.docx"],
    "instructions": "Extract all safety data and create a summary table showing adverse events by severity and frequency. Output as HTML with proper table formatting."
  }'

Administrators can make their custom agents and connectors visible to all users in the organization: Share Agent

POST /agents/{agentId}/share

Example Request

curl -X POST "https://api.artosai.com/agents/agent-user-custom-456/share" \
  -H "Authorization: Bearer admin-access-token" \
  -H "Content-Type: application/json" \
  -d '{"scope": "organization"}'

Share Connector

POST /connectors/{id}/share

Example Request

curl -X POST "https://api.artosai.com/connectors/conn-abc123def456/share" \
  -H "Authorization: Bearer admin-access-token" \
  -H "Content-Type: application/json" \
  -d '{"scope": "organization"}'

Visibility Rules

Regular Users: See only their own custom agents/connectors plus built-in/preset ones
Admin Users: See all agents/connectors in the organization
Shared Items: Custom items shared by admins are visible to all organization members

Error Handling

Common Error Codes

Error Code	HTTP Status	Description	Resolution
`CONNECTOR_NOT_FOUND`	404	Connector ID does not exist	Verify connector ID and permissions
`AGENT_NOT_FOUND`	404	Agent ID does not exist	Verify agent ID and availability
`INVALID_CONNECTOR_CONFIG`	400	Connector configuration is invalid	Check authentication credentials and config format
`AGENT_EXECUTION_FAILED`	500	Agent failed during execution	Check agent instructions and input documents
`MISSING_CREDENTIALS`	401	Connector authentication failed	Verify and update authentication credentials
`INSUFFICIENT_PERMISSIONS`	403	User lacks permission for operation	Contact admin or check user permissions
`INVALID_DOCUMENT_URL`	400	Document URL is not accessible	Ensure document is properly ingested
`CUSTOM_AGENT_LIMIT_EXCEEDED`	400	Too many custom agents (deprecated - no limits)	N/A
`CONNECTOR_IN_USE`	409	Cannot delete connector used by active pipelines	Remove connector from pipelines first

Error Response Format

{
  "error_code": "INVALID_CONNECTOR_CONFIG",
  "message": "The connector configuration contains invalid authentication parameters",
  "details": {
    "field": "config.client_secret",
    "issue": "Missing required OAuth client secret",
    "valid_auth_types": ["oauth", "api_key", "basic"]
  },
  "resolution": "Provide valid OAuth credentials or change to a different authentication type",
  "request_id": "req-abc123def456"
}

Specific Error Scenarios

Invalid Connector ID

{
  "error_code": "CONNECTOR_NOT_FOUND",
  "message": "The specified connector could not be found",
  "details": {
    "connector_id": "conn-nonexistent",
    "user_id": "user-xyz789"
  },
  "resolution": "Verify the connector ID exists and you have permission to access it"
}

Agent Execution Failure

{
  "error_code": "AGENT_EXECUTION_FAILED", 
  "message": "Agent failed to process the provided documents",
  "details": {
    "agent_id": "agent-custom-001",
    "execution_id": "exec-failed-123",
    "error_stage": "document_processing",
    "error_details": "Unable to parse document structure"
  },
  "resolution": "Check document format and agent instructions. Consider using a different agent or simplifying the task."
}

Missing Credentials

{
  "error_code": "MISSING_CREDENTIALS",
  "message": "Connector authentication failed due to missing or invalid credentials",
  "details": {
    "connector_id": "conn-abc123",
    "auth_type": "oauth",
    "missing_fields": ["client_secret"]
  },
  "resolution": "Update connector configuration with valid OAuth credentials"
}

Best Practices

Connector Management

Use descriptive names that clearly indicate the connector’s purpose
Test authentication before using connectors in production pipelines
Keep credentials secure and rotate them regularly
Document functions and rules thoroughly for team collaboration
Monitor usage to identify unused connectors

Agent Development

# Example: Testing custom agents before pipeline integration
import requests

class AgentTester:
    def __init__(self, base_url, access_token):
        self.base_url = base_url
        self.access_token = access_token
        self.headers = {
            'Authorization': f'Bearer {access_token}',
            'Content-Type': 'application/json'
        }
    
    def test_custom_agent(self, agent_config, test_documents):
        """Test custom agent with sample documents"""
        try:
            # Create agent
            response = requests.post(
                f"{self.base_url}/agents",
                headers=self.headers,
                json=agent_config
            )
            response.raise_for_status()
            agent_id = response.json()['id']
            
            # Test with documents
            test_response = requests.post(
                f"{self.base_url}/agents/{agent_id}/call",
                headers=self.headers,
                json={"documentUrls": test_documents}
            )
            test_response.raise_for_status()
            
            result = test_response.json()
            print(f"Agent test successful: {result['result']['output'][:100]}...")
            
            return agent_id, result
            
        except requests.RequestException as e:
            print(f"Agent test failed: {e}")
            raise
    
    def cleanup_test_agent(self, agent_id):
        """Remove test agent after validation"""
        try:
            requests.delete(
                f"{self.base_url}/agents/{agent_id}",
                headers=self.headers
            )
            print(f"Test agent {agent_id} cleaned up")
        except Exception as e:
            print(f"Cleanup warning: {e}")

# Usage
tester = AgentTester("https://api.artosai.com", "your-token")

agent_config = {
    "name": "Test Safety Analyzer",
    "description": "Test agent for safety data analysis",
    "instructions": "Extract and summarize safety information from clinical documents",
    "output_format": "html"
}

test_docs = ["s3://bucket/test-doc.docx"]
agent_id, result = tester.test_custom_agent(agent_config, test_docs)

# If successful, use in production; otherwise cleanup
if result['result']['output']:
    print("Agent ready for production use")
else:
    tester.cleanup_test_agent(agent_id)

Integration Patterns

Agent-Connector Workflow

// Example: Building a complete processing workflow
class DocumentProcessor {
    constructor(apiClient) {
        this.api = apiClient;
    }
    
    async setupProcessingWorkflow() {
        // 1. Create specialized connector
        const connector = await this.api.createConnector({
            name: "ClinicalDataConnector",
            sourceType: "database",
            config: {
                auth_type: "oauth",
                client_id: "clinical_app_id",
                client_secret: "clinical_secret",
                token_url: "https://clinical.company.com/oauth/token"
            },
            functions: [
                "extract_patient_demographics",
                "map_adverse_events",
                "validate_protocol_compliance"
            ]
        });
        
        // 2. Create custom analysis agent
        const agent = await this.api.createAgent({
            name: "Clinical Data Analyzer",
            description: "Analyzes clinical trial data with regulatory focus",
            instructions: `
                Analyze clinical trial data to generate regulatory-compliant summaries.
                Focus on:
                1. Patient demographics and baseline characteristics
                2. Efficacy outcomes and statistical significance
                3. Safety profile and adverse event analysis
                4. Protocol deviations and compliance
                
                Format output as structured HTML with clear sections and tables.
            `,
            output_format: "html"
        });
        
        // 3. Create pipeline combining them
        const pipeline = await this.api.createPipeline({
            name: "Clinical Analysis Pipeline",
            agentIds: ["agent-search-001", agent.id, "agent-qc-003"],
            order: [1, 2, 3],
            id: connector.id,
            documents: ["clinical-protocol-001", "patient-data-002"],
            outputFileName: "clinical-analysis-report.docx"
        });
        
        return {
            id: connector.id,
            agentId: agent.id,
            pipelineId: pipeline.pipelineId
        };
    }
}

Performance Optimization

Connector Reuse: Design connectors to be reusable across multiple agents and pipelines
Agent Specificity: Create focused agents for specific tasks rather than generic ones
Testing Strategy: Always test agents individually before using in pipelines
Error Handling: Implement robust error handling and retry logic
Monitoring: Track agent and connector performance for optimization opportunities

Rate Limiting

Rate limits apply at the LLM level for agent execution: Agent Execution Limits

Direct agent calls: Subject to LLM provider rate limits
Pipeline execution: Rate limited by the underlying AI models
Custom agent creation: No specific limits

Rate Limit Headers

X-RateLimit-Limit: 100
X-RateLimit-Remaining: 95
X-RateLimit-Reset: 1640995200
X-RateLimit-Type: llm-tokens

Analytics Integration

Agent and connector usage is tracked for analytics purposes. Detailed analytics documentation is available in a separate API documentation page. Tracked Metrics Include:

Agent execution frequency and success rates
Connector usage patterns and performance
Error rates and failure patterns
Processing time and resource utilization
User adoption and feature usage

API Documentation

Concepts

SDK Reference

Integrations

Deployment

Overview

Agents

Connectors

Connector Management

Create Connector

List All Connectors

Get Specific Connector

Update Connector

Delete Connector

Test Connector

Get Specific Agent

Create Custom Agent

Update Agent

Delete Agent

Call Agent Directly

CustomAgent Usage

Using CustomAgent in Pipelines

CustomAgent Direct Call

Visibility Rules

Error Handling

Common Error Codes

Error Response Format

Specific Error Scenarios

Best Practices

Connector Management

Agent Development

Integration Patterns

Performance Optimization

Rate Limiting

Analytics Integration

API Documentation

Concepts

SDK Reference

Integrations

Deployment

​Overview

​Agents

​Connectors

​Connector Management

​Create Connector

​List All Connectors

​Get Specific Connector

​Update Connector

​Delete Connector

​Test Connector

​Get Specific Agent

​Create Custom Agent

​Update Agent

​Delete Agent

​Call Agent Directly

​CustomAgent Usage

​Using CustomAgent in Pipelines

​CustomAgent Direct Call

​Permission-Based Sharing

​Admin Sharing (Admin Users Only)

​Visibility Rules

​Error Handling

​Common Error Codes

​Error Response Format

​Specific Error Scenarios

​Best Practices

​Connector Management

​Agent Development

​Integration Patterns

​Performance Optimization

​Rate Limiting

​Analytics Integration

Overview

Agents

Connectors

Connector Management

Create Connector

List All Connectors

Get Specific Connector

Update Connector

Delete Connector

Test Connector

Get Specific Agent

Create Custom Agent

Update Agent

Delete Agent

Call Agent Directly

CustomAgent Usage

Using CustomAgent in Pipelines

CustomAgent Direct Call

Permission-Based Sharing

Admin Sharing (Admin Users Only)

Visibility Rules

Error Handling

Common Error Codes

Error Response Format

Specific Error Scenarios

Best Practices

Connector Management

Agent Development

Integration Patterns

Performance Optimization

Rate Limiting

Analytics Integration