Deploy your AI workflows as production-ready REST APIs with one click. No server setup, no DevOps required. Automatic authentication, monitoring, and scaling included.

Prerequisites for Deployment

To deploy a flow as an API, it must have:

• ✅ API Input node - Defines the input schema for API requests
• ✅ At least one processing node - Prompts, web scrapers, or transformations
• ✅ Saved flow - Must be saved to the flow library

How to Deploy

1. 1

   Save Your Flow Click Save As in the Flow Playground and give your flow a descriptive title

2. 2

   Add API Input Node If not already present, add an API Input node at the start of your flow and configure input fields

3. 3

   Go to Saved Flows Navigate to Saved Flows in the sidebar

4. 4

   Select Your Flow Find and click on the flow you want to deploy

5. 5

   Click Deploy API Click the Deploy API button. Deployment takes 2-3 seconds.

6. 6

   Get API Endpoint Once deployed, you'll receive your unique API endpoint URL

API Endpoint Structure

Your deployed flow gets a unique endpoint:

https://api.evaligo.com/public/v1/flows/execute/{flow_id}

Where {flow_id} is your flow's unique identifier.

Making API Requests

Authentication

All API requests require authentication with your API key:

curl -X POST "https://api.evaligo.com/public/v1/flows/execute/abc123" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com",
    "query": "your data"
  }'

Request Body

The request body must match the fields defined in your API Input node:

{
  "field1": "value1",
  "field2": "value2"
}

Response Format

API calls return immediately with a job ID (async execution):

{
  "job_id": "job_abc123xyz",
  "status": "queued",
  "flow_id": "abc123"
}

Async Execution Model

All flow executions are asynchronous to handle long-running operations:

1. Submit Request

Your API call returns immediately with a job ID

2. Poll for Results

Use the job status endpoint to check progress:

curl "https://api.evaligo.com/public/v1/jobs/{job_id}" \
  -H "Authorization: Bearer YOUR_API_KEY"

3. Get Results

When status is completed, the response includes results:

{
  "job_id": "job_abc123xyz",
  "status": "completed",
  "flow_id": "abc123",
  "result": {
    "output": "Your flow result data here"
  },
  "execution_time_seconds": 5.2,
  "completed_at": "2025-10-17T12:34:56Z"
}

Code Examples

See complete working examples in multiple languages:

Monitoring & Observability

Monitor your deployed API:

• Execution Jobs Table - View all API requests and their status
• Cost Tracking - See API usage costs in real-time
• Error Logs - Failed requests with detailed error messages
• Execution Time - Performance metrics per endpoint

Rate Limits & Scaling

Concurrency Limits

• Free tier: 4 concurrent executions
• Pro tier: 20 concurrent executions
• Enterprise: Custom limits

Request Limits

• Free tier: 1,000 requests/month
• Pro tier: 10,000 requests/month
• Enterprise: Unlimited

Undeploying APIs

To stop accepting API requests:

1. Go to Saved Flows
2. Select the deployed flow
3. Click Undeploy API

Best Practices

Security

• ✅ Rotate API keys regularly
• ✅ Use separate API keys for dev/staging/production
• ✅ Never commit API keys to version control
• ✅ Use environment variables for API keys

Error Handling

• ✅ Implement retry logic with exponential backoff
• ✅ Handle 429 (rate limit) errors gracefully
• ✅ Log failed requests for debugging
• ✅ Set timeouts on your HTTP client (300 seconds recommended)

Testing

• ✅ Test thoroughly in Flow Playground before deploying
• ✅ Use small batches to test API behavior
• ✅ Monitor first few production requests closely
• ✅ Have rollback plan (keep previous version deployable)

Troubleshooting

401 Unauthorized

• Check API key is correct
• Verify Authorization: Bearer YOUR_KEY header format
• Ensure API key hasn't been revoked

400 Bad Request

• Request body doesn't match API Input node schema
• Missing required fields
• Invalid field types (e.g., string instead of number)

429 Too Many Requests

• You've exceeded concurrent execution limit
• Wait for current jobs to complete
• Implement queuing on your side
• Consider upgrading your plan

Job stays in "processing" forever

• Check if flow has infinite loops (not supported yet)
• Verify all nodes are working correctly
• Contact support if issue persists