Deploy Flow - GraphorLM Docs

Deploy your configured flows to make them publicly accessible and executable via the GraphorLM REST API. Deployment creates a new revision and makes your flow available at a public URL.

Overview

The Deploy Flow endpoint creates a new revision of your flow and deploys it, making it accessible for execution via the public API. Once deployed, your flow will be available at https://{flow_name}.flows.graphorlm.com.

Response Node Required: Your flow must have a configured Response Node for deployment to work properly. The deployment process validates that all required components, including the Response Node, are properly configured.

Method: POST
URL: https://{flow_name}.flows.graphorlm.com/deploy
Authentication: Required (API Token)

Authentication

All requests must include a valid API token in the Authorization header:

Authorization: Bearer YOUR_API_TOKEN

Learn how to generate API tokens in the API Tokens guide.

Request Format

Headers

Header	Value	Required
`Authorization`	`Bearer YOUR_API_TOKEN`	Yes
`Content-Type`	`application/json`	Yes

Request Body

The request body should be a JSON object with the following optional fields:

Field	Type	Required	Description
`tool_description`	string	No	Custom description for the deployed flow that will be used in tool definitions

Example Request

{
  "tool_description": "Advanced RAG pipeline for technical documentation with semantic search and context-aware responses"
}

Minimal Request

For a basic deployment without custom tool description:

{}

Response Format

Success Response (200 OK)

{
  "success": true,
  "message": "Flow 'my-rag-pipeline' deployed successfully",
  "revision_id": "550e8400-e29b-41d4-a716-446655440000",
  "status": "Deployed"
}

Response Fields

Field	Type	Description
`success`	boolean	Whether the deployment was successful
`message`	string	Human-readable message about the deployment
`revision_id`	string	Unique identifier for the created revision
`status`	string	Current status of the flow (should be “Deployed”)

Code Examples

JavaScript/Node.js

async function deployFlow(flowName, toolDescription = null) {
  const url = `https://${flowName}.flows.graphorlm.com/deploy`;
  
  const payload = {};
  if (toolDescription) {
    payload.tool_description = toolDescription;
  }

  const response = await fetch(url, {
    method: 'POST',
    headers: {
      'Authorization': 'Bearer YOUR_API_TOKEN',
      'Content-Type': 'application/json',
    },
    body: JSON.stringify(payload)
  });

  if (!response.ok) {
    const errorData = await response.json();
    throw new Error(`Deployment failed: ${errorData.detail}`);
  }

  return await response.json();
}

// Usage
deployFlow('my-rag-pipeline', 'Advanced document Q&A system')
  .then(result => {
    console.log('✅ Deployment successful!');
    console.log('Revision ID:', result.revision_id);
    console.log('Status:', result.status);
    console.log('Message:', result.message);
  })
  .catch(error => {
    console.error('❌ Deployment failed:', error.message);
  });

// Simple deployment without custom description
deployFlow('my-rag-pipeline')
  .then(result => console.log('Deployed:', result))
  .catch(error => console.error('Error:', error));

Python

import requests
import json

def deploy_flow(flow_name, api_token, tool_description=None):
    url = f"https://{flow_name}.flows.graphorlm.com/deploy"
    
    headers = {
        "Authorization": f"Bearer {api_token}",
        "Content-Type": "application/json"
    }
    
    payload = {}
    if tool_description:
        payload["tool_description"] = tool_description
    
    response = requests.post(url, headers=headers, json=payload)
    response.raise_for_status()
    
    return response.json()

# Usage
try:
    result = deploy_flow(
        flow_name="my-rag-pipeline",
        api_token="YOUR_API_TOKEN",
        tool_description="Technical documentation assistant with advanced search capabilities"
    )
    
    print("✅ Deployment successful!")
    print(f"Revision ID: {result['revision_id']}")
    print(f"Status: {result['status']}")
    print(f"Message: {result['message']}")
    
    # Flow is now available at:
    flow_url = f"https://{result.get('flow_name', 'my-rag-pipeline')}.flows.graphorlm.com"
    print(f"🚀 Flow available at: {flow_url}")
    
except requests.exceptions.HTTPError as e:
    if e.response.status_code == 404:
        print("❌ Flow not found. Make sure the flow exists and is properly configured.")
    elif e.response.status_code == 400:
        print("❌ Invalid flow configuration. Check your flow setup.")
    else:
        print(f"❌ Deployment failed: {e}")
        try:
            error_detail = e.response.json().get('detail', 'Unknown error')
            print(f"Error details: {error_detail}")
        except:
            print(f"HTTP {e.response.status_code}: {e.response.text}")

except Exception as e:
    print(f"❌ Unexpected error: {e}")

cURL

# Deploy with custom tool description
curl -X POST https://my-rag-pipeline.flows.graphorlm.com/deploy \
  -H "Authorization: Bearer YOUR_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "tool_description": "Advanced RAG pipeline for technical documentation"
  }'

# Simple deployment without custom description
curl -X POST https://my-rag-pipeline.flows.graphorlm.com/deploy \
  -H "Authorization: Bearer YOUR_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{}'

PHP

<?php
function deployFlow($flowName, $apiToken, $toolDescription = null) {
    $url = "https://{$flowName}.flows.graphorlm.com/deploy";
    
    $payload = [];
    if ($toolDescription) {
        $payload['tool_description'] = $toolDescription;
    }
    
    $options = [
        'http' => [
            'header' => [
                "Authorization: Bearer {$apiToken}",
                "Content-Type: application/json"
            ],
            'method' => 'POST',
            'content' => json_encode($payload)
        ]
    ];
    
    $context = stream_context_create($options);
    $result = file_get_contents($url, false, $context);
    
    if ($result === FALSE) {
        $error = error_get_last();
        throw new Exception("Failed to deploy flow: " . $error['message']);
    }
    
    $httpCode = null;
    if (isset($http_response_header)) {
        foreach ($http_response_header as $header) {
            if (strpos($header, 'HTTP/') === 0) {
                $httpCode = (int) substr($header, 9, 3);
                break;
            }
        }
    }
    
    $response = json_decode($result, true);
    
    if ($httpCode !== 200) {
        $errorMessage = $response['detail'] ?? 'Unknown error';
        throw new Exception("Deployment failed: {$errorMessage}");
    }
    
    return $response;
}

// Usage
try {
    $result = deployFlow(
        'my-rag-pipeline',
        'YOUR_API_TOKEN',
        'Advanced document analysis and Q&A system'
    );
    
    echo "✅ Deployment successful!\n";
    echo "Revision ID: {$result['revision_id']}\n";
    echo "Status: {$result['status']}\n";
    echo "Message: {$result['message']}\n";
    
    $flowUrl = "https://my-rag-pipeline.flows.graphorlm.com";
    echo "🚀 Flow available at: {$flowUrl}\n";
    
} catch (Exception $e) {
    echo "❌ Deployment failed: " . $e->getMessage() . "\n";
}
?>

Deployment Process

The deployment process involves several steps:

Validation: Checks if the flow exists and is properly configured
Revision Creation: Creates a new revision with updated node IDs
Graph Update: Updates the flow graph with the new revision
Component Deployment: Deploys all flow components (dataset, chunking, retrieval, Response Node, etc.)
Traffic Routing: Routes 100% of traffic to the new revision
Status Update: Updates the flow status to “Deployed”

Deployment can take several minutes depending on flow complexity and the number of components that need to be processed.

Deployment Requirements

Before deploying a flow, ensure:

1. Flow Configuration

✅ Flow has been created and saved
✅ Response Node is configured and properly connected
✅ All required components are configured
✅ Dataset node has source documents selected
✅ Response Node model and prompt configuration is complete and valid

2. Source Documents

✅ Documents are uploaded to the project
✅ Documents are processed and in “Completed” status
✅ At least one document is selected in dataset nodes

3. Flow Components

✅ All nodes have valid configurations
✅ Node connections are properly established
✅ Required parameters are set for each component

4. API Access

✅ Valid API token with deployment permissions
✅ Flow name is unique and follows naming conventions

Error Responses

Common Error Codes

Status Code	Description	Common Causes
400	Bad Request	Invalid flow configuration, missing required components
401	Unauthorized	Invalid or missing API token
404	Not Found	Flow doesn’t exist or insufficient permissions
500	Internal Server Error	Deployment process failed, infrastructure issues

Error Response Format

{
  "detail": "Error message describing what went wrong"
}

Example Error Responses

Flow Not Found

{
  "detail": "Flow with name 'my-rag-pipeline' not found"
}

Invalid Configuration

{
  "detail": "Flow configuration is invalid: Dataset node is missing source documents"
}

Deployment Failed

{
  "detail": "Failed to deploy flow: Component validation failed"
}

Missing Response Node

{
  "detail": "Failed to deploy flow: No Response Node found in flow"
}

Insufficient Permissions

{
  "detail": "You don't have permission to deploy flows in this project"
}

Tool Description Usage

The tool_description parameter is used when your deployed flow is accessed via MCP (Model Context Protocol) or when it’s used as a tool in other systems. It helps describe what your flow does and how it should be used.

Best Practices for Tool Descriptions

Be Specific: Clearly describe what your flow does
Include Use Cases: Mention what types of queries it handles best
Mention Data Sources: Indicate what kind of documents or data it processes

Examples of Good Tool Descriptions

{
  "tool_description": "Technical documentation Q&A system that searches through software manuals, API docs, and troubleshooting guides to provide accurate answers to development questions"
}

{
  "tool_description": "Legal document analysis tool that processes contracts, regulations, and case studies to answer legal research questions with cited sources"
}

{
  "tool_description": "Customer support knowledge base that searches through product manuals, FAQ documents, and support tickets to provide helpful customer service responses"
}

Integration Examples

Automated Deployment Pipeline

class FlowDeploymentManager {
  constructor(apiToken) {
    this.apiToken = apiToken;
  }

  async deployWithValidation(flowName, toolDescription) {
    try {
      // Step 1: Validate flow configuration
      console.log('📋 Validating flow configuration...');
      await this.validateFlowConfiguration(flowName);
      
      // Step 2: Deploy the flow
      console.log('🚀 Starting deployment...');
      const result = await this.deployFlow(flowName, toolDescription);
      
      // Step 3: Verify deployment
      console.log('✅ Verifying deployment...');
      await this.verifyDeployment(flowName);
      
      console.log('🎉 Deployment completed successfully!');
      return result;
      
    } catch (error) {
      console.error('❌ Deployment failed:', error.message);
      throw error;
    }
  }

  async deployFlow(flowName, toolDescription) {
    const url = `https://${flowName}.flows.graphorlm.com/deploy`;
    
    const payload = {};
    if (toolDescription) {
      payload.tool_description = toolDescription;
    }

    const response = await fetch(url, {
      method: 'POST',
      headers: {
        'Authorization': `Bearer ${this.apiToken}`,
        'Content-Type': 'application/json',
      },
      body: JSON.stringify(payload)
    });

    if (!response.ok) {
      const errorData = await response.json();
      throw new Error(`Deployment failed: ${errorData.detail}`);
    }

    return await response.json();
  }

  async validateFlowConfiguration(flowName) {
    // Check if flow exists and is properly configured
    // This would typically involve checking dataset nodes, Response Node, etc.
    const datasetsUrl = `https://${flowName}.flows.graphorlm.com/datasets`;
    
    const response = await fetch(datasetsUrl, {
      headers: {
        'Authorization': `Bearer ${this.apiToken}`,
      }
    });

    if (!response.ok) {
      throw new Error('Flow configuration validation failed');
    }

    const datasets = await response.json();
    if (datasets.length === 0) {
      throw new Error('Flow has no dataset nodes configured');
    }

    // Additional validation logic here
    // You would also check for Response Node presence:
    // const llmUrl = `https://${flowName}.flows.graphorlm.com/llm`;
    // const llmResponse = await fetch(llmUrl, { headers: { 'Authorization': `Bearer ${this.apiToken}` }});
    // const llmNodes = await llmResponse.json();
    // if (llmNodes.length === 0) {
    //   throw new Error('Flow has no Response Node configured');
    // }
    
    console.log('✅ Flow configuration is valid');
  }

  async verifyDeployment(flowName) {
    // Test the deployed flow with a simple query
    const testUrl = `https://${flowName}.flows.graphorlm.com`;
    
    const response = await fetch(testUrl, {
      method: 'POST',
      headers: {
        'Authorization': `Bearer ${this.apiToken}`,
        'Content-Type': 'application/json',
      },
      body: JSON.stringify({
        query: 'test query',
        page_size: 1
      })
    });

    if (!response.ok) {
      throw new Error('Deployment verification failed - flow is not responding');
    }

    console.log('✅ Flow is responding correctly');
  }
}

// Usage
const deploymentManager = new FlowDeploymentManager('YOUR_API_TOKEN');

deploymentManager.deployWithValidation(
  'my-rag-pipeline',
  'Advanced technical documentation assistant'
).then(result => {
  console.log('Deployment result:', result);
}).catch(error => {
  console.error('Deployment error:', error);
});

Batch Deployment for Multiple Flows

import asyncio
import aiohttp
import json
from typing import List, Dict, Any

class BatchFlowDeployer:
    def __init__(self, api_token: str):
        self.api_token = api_token
        
    async def deploy_single_flow(
        self, 
        session: aiohttp.ClientSession, 
        flow_name: str, 
        tool_description: str = None
    ) -> Dict[str, Any]:
        url = f"https://{flow_name}.flows.graphorlm.com/deploy"
        
        payload = {}
        if tool_description:
            payload["tool_description"] = tool_description
            
        headers = {
            "Authorization": f"Bearer {self.api_token}",
            "Content-Type": "application/json"
        }
        
        try:
            async with session.post(url, headers=headers, json=payload) as response:
                if response.status == 200:
                    result = await response.json()
                    return {
                        "flow_name": flow_name,
                        "status": "success",
                        "revision_id": result["revision_id"],
                        "message": result["message"]
                    }
                else:
                    error_text = await response.text()
                    return {
                        "flow_name": flow_name,
                        "status": "error",
                        "error": error_text
                    }
        except Exception as e:
            return {
                "flow_name": flow_name,
                "status": "error",
                "error": str(e)
            }
    
    async def deploy_multiple_flows(
        self, 
        flows: List[Dict[str, str]]
    ) -> List[Dict[str, Any]]:
        """
        Deploy multiple flows concurrently.
        
        Args:
            flows: List of dicts with 'name' and optional 'tool_description'
        """
        async with aiohttp.ClientSession() as session:
            tasks = [
                self.deploy_single_flow(
                    session, 
                    flow["name"], 
                    flow.get("tool_description")
                )
                for flow in flows
            ]
            return await asyncio.gather(*tasks, return_exceptions=True)

# Usage
async def deploy_all_flows():
    deployer = BatchFlowDeployer("YOUR_API_TOKEN")
    
    flows_to_deploy = [
        {
            "name": "customer-support-flow",
            "tool_description": "Customer support Q&A system"
        },
        {
            "name": "technical-docs-flow", 
            "tool_description": "Technical documentation assistant"
        },
        {
            "name": "legal-research-flow",
            "tool_description": "Legal document research tool"
        }
    ]
    
    print("🚀 Starting batch deployment...")
    results = await deployer.deploy_multiple_flows(flows_to_deploy)
    
    successful = 0
    failed = 0
    
    for result in results:
        if isinstance(result, dict):
            if result["status"] == "success":
                print(f"✅ {result['flow_name']}: {result['message']}")
                successful += 1
            else:
                print(f"❌ {result['flow_name']}: {result['error']}")
                failed += 1
        else:
            print(f"❌ Unexpected error: {result}")
            failed += 1
    
    print(f"\n📊 Deployment Summary:")
    print(f"✅ Successful: {successful}")
    print(f"❌ Failed: {failed}")

# Run batch deployment
# asyncio.run(deploy_all_flows())

Best Practices

Pre-Deployment Checklist

✅ Ensure Response Node is configured and connected
✅ Test your flow configuration in the GraphorLM interface
✅ Verify all documents are processed and available
✅ Ensure dataset nodes have appropriate source selections
✅ Test component configurations individually
✅ Validate Response Node model, prompt, and API settings

Deployment Strategy

Start Small: Deploy simple flows first to validate the process
Monitor Performance: Watch deployment progress and check for errors
Version Control: Keep track of revision IDs for rollback if needed
Gradual Rollout: Test thoroughly before production deployment

Post-Deployment

Verify Functionality: Test the deployed flow with sample queries
Monitor Performance: Check response times and accuracy
Update Documentation: Document the deployment and any custom configurations
Set Up Monitoring: Implement logging and error tracking

Security Considerations

API Token Management: Use environment variables for tokens
Access Control: Ensure only authorized users can deploy flows
Network Security: Use HTTPS for all deployment requests
Audit Trail: Keep logs of all deployments for compliance

Troubleshooting

Flow Not Found Error

Invalid Configuration Error

Missing Response Node

Deployment Timeout

Source Document Issues

Authentication Failures

Internal Server Errors

Monitoring and Management

After successful deployment, you can:

Check Deployment Status

Monitor flow execution through the GraphorLM dashboard
Use the Run Flow endpoint to test functionality
Review logs and performance metrics

Update Deployments

Modify your flow configuration in the GraphorLM interface
Deploy again to create a new revision
Previous revisions are automatically replaced

Rollback if Needed

Contact support if you need to rollback to a previous revision
Keep track of working revision IDs for reference

Next Steps

After successfully deploying your flow, you can:

Run Flow

Execute your deployed flow with queries and get responses

Flow Overview

Explore all available flow management endpoints

Integrate Workflow

Advanced integration patterns and MCP server usage

API Tokens

Learn how to generate and manage API tokens

Get Started

Sources

Flows

​Overview

​Authentication

​Request Format

​Headers

​Request Body

​Example Request

​Minimal Request

​Response Format

​Success Response (200 OK)

​Response Fields

​Code Examples

​JavaScript/Node.js

​Python

​cURL

​PHP

​Deployment Process

​Deployment Requirements

​1. Flow Configuration

​2. Source Documents

​3. Flow Components

​4. API Access

​Error Responses

​Common Error Codes

​Error Response Format

​Example Error Responses

​Flow Not Found

​Invalid Configuration

​Deployment Failed

​Missing Response Node

​Insufficient Permissions

​Tool Description Usage

​Best Practices for Tool Descriptions

​Examples of Good Tool Descriptions

​Integration Examples

​Automated Deployment Pipeline

​Batch Deployment for Multiple Flows

​Best Practices

​Pre-Deployment Checklist

​Deployment Strategy

​Post-Deployment

​Security Considerations

​Troubleshooting

​Monitoring and Management

​Check Deployment Status

​Update Deployments

​Rollback if Needed

​Next Steps

Run Flow

Flow Overview

Integrate Workflow

API Tokens

Overview

Authentication

Request Format

Headers

Request Body

Example Request

Minimal Request

Response Format

Success Response (200 OK)

Response Fields

Code Examples

JavaScript/Node.js

Python

cURL

PHP

Deployment Process

Deployment Requirements

1. Flow Configuration

2. Source Documents

3. Flow Components

4. API Access

Error Responses

Common Error Codes

Error Response Format

Example Error Responses

Flow Not Found

Invalid Configuration

Deployment Failed

Missing Response Node

Insufficient Permissions

Tool Description Usage

Best Practices for Tool Descriptions

Examples of Good Tool Descriptions

Integration Examples

Automated Deployment Pipeline

Batch Deployment for Multiple Flows

Best Practices

Pre-Deployment Checklist

Deployment Strategy

Post-Deployment

Security Considerations

Troubleshooting

Monitoring and Management

Check Deployment Status

Update Deployments

Rollback if Needed

Next Steps