HTTP/HTTPS API

Connect to any HTTP API endpoint with full control over request/response handling.

Configuration

Field	Required	Description
Application Name	Yes	Unique identifier (e.g., my-http-app)
API Endpoint URL	Yes	Full URL to the API endpoint
HTTP Method	No	POST (default) or GET
HTTP Headers	No	Custom headers as key-value pairs
Request Body	Yes	JSON template with placeholders
Response Parser Path	Yes	JSONPath to extract the response

Available Placeholders

Use these placeholders in your request body:

Placeholder	Description	Example
{{messages}}	Full conversation as OpenAI-format array	[{"role": "user", "content": "Hello"}]
{{prompt}}	Formatted conversation as string	"User: Hello\n\nAssistant: Hi!"

Response Parser Syntax

Extract the AI response using dot notation and array indexing:

Parser Path	Use Case
choices[0].message.content	OpenAI-style responses
response.text	Simple nested object
data[0].output	Array response
content	Top-level field
result.candidates[0].text	Nested array

Example Configurations

OpenAI-Style API

Endpoint: https://api.example.com/v1/chat/completions

Headers:

Key	Value
Content-Type	application/json
Authorization	Bearer YOUR_API_KEY

Request Body:

{
  "model": "gpt-3.5-turbo",
  "messages": {{messages}},
  "temperature": 0.7
}

Response Parser Path:

choices[0].message.content

Simple Text API

Endpoint: https://api.example.com/generate

Headers:

Key	Value
X-API-Key	YOUR_API_KEY

Request Body:

{
  "text": "{{prompt}}",
  "max_length": 100
}

Response Parser Path:

generated_text

Custom Inference Server

Endpoint: https://my-server.com/inference

Headers:

Key	Value
Content-Type	application/json

Request Body:

{
  "inputs": "{{prompt}}",
  "parameters": {
    "max_new_tokens": 512,
    "temperature": 0.8
  }
}

Response Parser Path:

output

Troubleshooting

Issue	Solution
Empty response	Verify the Response Parser Path matches your API's response structure
401 Unauthorized	Check API key in headers
400 Bad Request	Validate JSON structure in Request Body
Timeout	Ensure endpoint is reachable and responding

Setup Steps

1. Navigate to AI Applications → New Application
2. Select Custom Applications tab
3. Click HTTP/HTTPS API
4. Enter your Application Name
5. Enter your API Endpoint URL
6. Select HTTP Method (POST or GET)
7. Add HTTP Headers (e.g., Authorization, Content-Type)
8. Enter your Request Body template with placeholders
9. Enter your Response Parser Path
10. Click Test Response to verify (optional)
11. Review and submit