Endpoints

Overview

Endpoints allow you to create custom API interfaces with secure, controlled access. You can define three types of endpoints:

Query Endpoints - Execute SQL queries on your datasets with dynamic filters
Action Endpoints - Execute serverless Python functions for custom logic
Vector Endpoints - Perform semantic similarity search on vectorized data

Creating Endpoints

Navigate to your Project
Create a new YAML file for your endpoint (e.g., api-orders.yaml)
Define your endpoint configuration
Reference it in your config.yaml file

Endpoint Types

Although the YAML syntax uses an endpoints: array, only one endpoint definition per file is supported. Defining multiple entries in a single file will result in an error. Use a separate YAML file for each endpoint and register each file individually in config.yaml.

Query-Based Endpoints

Query endpoints execute SQL queries on your data with dynamic filtering using Jinja templating. Example YAML Configuration:

endpoints:
  - name: consolidated-sales-endpoint
    type: query
    config:
      filters:
        - name: country
          type: string
          optional: true
        - name: city
          type: string
          optional: true
      query: |
          SELECT *
          FROM consolidated_sales_df_299ceb
          WHERE 1 = 1
          {% if country is defined %}
          AND CustomerCountry = '{{ country }}'
          {% endif %}
          {% if city is defined %}
          AND CustomerCity = '{{ city }}'
          {% endif %}

Action-Based Endpoints

Action endpoints execute serverless Python functions when called. Perfect for sending notifications, processing data, calling external APIs, or automating workflows. Example YAML Configuration:

endpoints:
  - name: process-data-endpoint
    type: action
    config:
      action_id: "507f1f77bcf86cd799439011"

The action_id references an action function in your project. Get the ID from your action details page.

The action function must return a list. Returning any other type raises ActionEndpointResultMustBeListError.

Vector-Based Endpoints

Vector endpoints enable semantic similarity search on your vectorized data via HTTP API. Perfect for building search features, recommendation systems, or RAG applications. Example YAML Configuration:

endpoints:
  - name: document-search-endpoint
    type: vector
    config:
      vector_id: "69d597a6e9713d78370a50d9"

The vector_id references a Vector in your project. Get the ID from your vector details page.

Register in config.yaml

Reference your endpoint file in config.yaml:

project_name: TestProject
project_id: 688b3d0b7c5f93f0763a028c
endpoints:
  - path: api-orders.yaml

Learn more about the configuration file in the Project Configuration section.

Configuration Reference

Common Attributes

Attribute	Type	Required	Description
`name`	string	Yes	Unique endpoint identifier
`type`	string	Yes	Endpoint type: `query`, `action`, or `vector`
`config`	object	Yes	Type-specific configuration

Query Config Attributes

Attribute	Type	Required	Description
`query`	string	Yes	SQL query with Jinja templating
`filters`	array	No	List of filter parameters

Action Config Attributes

Attribute	Type	Required	Description
`action_id`	string	Yes	Action function ID to execute

Vector Config Attributes

Attribute	Type	Required	Description
`vector_id`	string	Yes	Vector ID for similarity search

Filter Configuration

Filters are only for query endpoints and define dynamic parameters:

Attribute	Type	Required	Description
`name`	string	Yes	Filter parameter name (alphanumeric, -, _)
`type`	string	Yes	Data type (see types below)
`optional`	boolean	No	Whether filter is optional (default: `true`)

Filter Types

Type	Description	Example Usage
`string`	Text parameter	`country`, `category`
`integer`	Numeric parameter	`year`, `limit`
`float`	Decimal parameter	`price`, `rating`
`date`	Date parameter	`start_date`, `end_date`
`datetime`	DateTime parameter	`created_at`, `updated_at`
`boolean`	Boolean parameter	`active`, `is_featured`
`enum`	Enumeration parameter	`status`, `priority`

Using Endpoints

Endpoints are accessible via HTTP requests. Query and vector endpoints use GET requests, while action endpoints may vary based on implementation:

curl -X GET \
  "https://app.datazone.co/api/v1/endpoint/your-endpoint-name?param1=value1&param2=value2" \
  -H "x-api-key: <DATAZONE_API_KEY>" \
  -H "Content-Type: application/json"

import requests

url = "https://app.datazone.co/api/v1/endpoint/your-endpoint-name"
headers = {
    "x-api-key": "<DATAZONE_API_KEY>",
    "Content-Type": "application/json"
}
params = {
    "param1": "value1",
    "param2": "value2"
}

response = requests.get(url, headers=headers, params=params)
data = response.json()
print(data)

const url = 'https://app.datazone.co/api/v1/endpoint/your-endpoint-name';
const params = new URLSearchParams({
    param1: 'value1',
    param2: 'value2'
});

const response = await fetch(`${url}?${params}`, {
    method: 'GET',
    headers: {
        'x-api-key': '<DATAZONE_API_KEY>',
        'Content-Type': 'application/json'
    }
});

const data = await response.json();
console.log(data);

import java.net.http.HttpClient;
import java.net.http.HttpRequest;
import java.net.http.HttpResponse;
import java.net.URI;

HttpClient client = HttpClient.newHttpClient();
String url = "https://app.datazone.co/api/v1/endpoint/your-endpoint-name?param1=value1&param2=value2";

HttpRequest request = HttpRequest.newBuilder()
    .uri(URI.create(url))
    .header("x-api-key", "<DATAZONE_API_KEY>")
    .header("Content-Type", "application/json")
    .GET()
    .build();

HttpResponse<String> response = client.send(request, HttpResponse.BodyHandlers.ofString());
System.out.println(response.body());

Response Format

Query Endpoint Response

Query endpoints return JSON data with query results:

{
  "records": [
    {
      "column1": "value1",
      "column2": "value2"
    }
  ],
  "metadata": {
    "total_records": 100,
    "query_time_ms": 250,
    "endpoint_name": "consolidated-sales-endpoint"
  }
}

Action Endpoint Response

Action endpoints return the list your Python function returns:

[
  {
    "status": "success",
    "message": "Action executed successfully"
  }
]

Your action function must return a list. Returning any other type (dict, string, None, etc.) raises ActionEndpointResultMustBeListError. Always wrap your result in a list, even when there is only one item.

# ✅ Correct — return a list
def run(context):
    return [{"status": "success", "message": "Done"}]

# ❌ Wrong — raises ActionEndpointResultMustBeListError
def run(context):
    return {"status": "success", "message": "Done"}

Vector Endpoint Response

Vector endpoints return semantically similar results based on your search query:

{
  "results": [
    {
      "content": "Relevant text chunk from your data...",
      "metadata": {
        "source": "document.pdf",
        "page": 5,
        "chunk_id": "abc123"
      },
      "score": 0.92
    },
    {
      "content": "Another relevant text chunk...",
      "metadata": {
        "source": "report.docx",
        "section": "Introduction"
      },
      "score": 0.87
    }
  ],
  "metadata": {
    "vector_id": "69d597a6e9713d78370a50d9",
    "endpoint_name": "document-search-endpoint",
    "query": "what is the company revenue?",
    "total_results": 10
  }
}

Using Vector Endpoints:

curl -X GET \
  "https://app.datazone.co/api/v1/endpoint/document-search-endpoint?query=what+is+the+company+revenue" \
  -H "x-api-key: <DATAZONE_API_KEY>" \
  -H "Content-Type: application/json"

import requests

url = "https://app.datazone.co/api/v1/endpoint/document-search-endpoint"
headers = {
    "x-api-key": "<DATAZONE_API_KEY>",
    "Content-Type": "application/json"
}
params = {
    "query": "what is the company revenue?"
}

response = requests.get(url, headers=headers, params=params)
results = response.json()

for result in results["results"]:
    print(f"Score: {result['score']}")
    print(f"Content: {result['content']}")
    print(f"Metadata: {result['metadata']}\n")

Query Parameters for Vector Endpoints:

Parameter	Type	Required	Description
`query`	string	Yes	Search query in natural language

Authentication

Endpoints use API key authentication via the x-api-key header:

-H "x-api-key: YOUR_API_KEY"

Error Responses

Status Code	Description
200	Success
400	Bad Request - Invalid parameters
401	Unauthorized - Invalid API key
403	Forbidden - Access denied
404	Not Found - Endpoint doesn’t exist
500	Internal Server Error

Example error response:

{
  "error": "Invalid parameter 'country': must be a string",
  "code": "INVALID_PARAMETER",
  "status": 400
}

Advanced Query Features

Jinja Templating

Endpoints support Jinja templating for dynamic queries:

query: |
    SELECT *
    FROM sales_data
    WHERE 1 = 1
    {% if start_date is defined %}
    AND order_date >= '{{ start_date }}'
    {% endif %}
    {% if end_date is defined %}
    AND order_date <= '{{ end_date }}'
    {% endif %}
    {% if categories is defined %}
    AND category IN ({{ categories | join("','") | surround("'") }})
    {% endif %}

Caching

Endpoint responses can be cached:

Default cache TTL: 60 minutes
Cache headers indicate freshness
Use cache-busting parameters when needed

Get Started

Reference

Deployment

Tutorial

Overview

Creating Endpoints

Endpoint Types

Query-Based Endpoints

Action-Based Endpoints

Vector-Based Endpoints

Register in config.yaml

Configuration Reference

Common Attributes

Query Config Attributes

Action Config Attributes

Vector Config Attributes

Filter Configuration

Filter Types

Using Endpoints

Response Format

Query Endpoint Response

Action Endpoint Response

Vector Endpoint Response

Authentication

Error Responses

Advanced Query Features

Jinja Templating

Caching

​Overview

​Creating Endpoints

​Endpoint Types

​Query-Based Endpoints

​Action-Based Endpoints

​Vector-Based Endpoints

​Register in config.yaml

​Configuration Reference

​Common Attributes

​Query Config Attributes

​Action Config Attributes

​Vector Config Attributes

​Filter Configuration

​Filter Types

​Using Endpoints

​Response Format

​Query Endpoint Response

​Action Endpoint Response

​Vector Endpoint Response

​Authentication

​Error Responses

​Advanced Query Features

​Jinja Templating

​Caching

Overview

Creating Endpoints

Endpoint Types

Query-Based Endpoints

Action-Based Endpoints

Vector-Based Endpoints

Register in config.yaml

Configuration Reference

Common Attributes

Query Config Attributes

Action Config Attributes

Vector Config Attributes

Filter Configuration

Filter Types

Using Endpoints

Response Format

Query Endpoint Response

Action Endpoint Response

Vector Endpoint Response

Authentication

Error Responses

Advanced Query Features

Jinja Templating

Caching