Documentation Index
Fetch the complete documentation index at: https://docs.datalinks.com/llms.txt
Use this file to discover all available pages before exploring further.
This guide shows how to connect a third-party web app builder (e.g. Lovable, Replit, Vercel v0) to DataLinks using a starter prompt and the DataLinks API. The same approach works for chat apps, dashboards, internal tools, and AI assistants.
What You Will Need
Before you start, make sure you have the following:
- A DataLinks API key
- Your namespace name
- One or more dataset names inside that namespace
- A web app builder that supports API calls or server-side functions
Step 1: Create a Clear Starter Prompt
Most app builders allow you to define an initial system or starter prompt. This prompt tells the app how it should behave and what data it can access.
Example starter prompt:
Create a simple chatbox interface that connects to the <DATASET_1> and <DATASET_2> datasets within the <NAMESPACE> namespace. The chatbox should perform AutoRAG queries on the data in DataLinks when queried.
- Clearly name the namespace and datasets
- State whether the app should use structured queries, natural language queries, or AutoRAG
- Describe the expected output format, such as a text response or table
In your web app builder, create a new API connection or backend function with the following settings.
Base URL
https://api.datalinks.com/api/v1
Authorization: Bearer <DL_API_KEY>
Step 3: Choose the Right Endpoint
Use the endpoint that matches how your app needs to interact with the data.
Option A: Structured or Natural Language Queries
Endpoint:
Use this when the app needs to query a specific dataset.
Example request body for a natural language query:
{
"username": "self",
"namespace": "<NAMESPACE>",
"dataset": "<OBJECT_NAME>",
"naturalLanguageQuery": "Show me the latest 50 records"
}
{
"data": [
{ "field1": "value", "field2": 123, "field3": true }
]
}
Option B: AutoRAG Question Answering
Endpoint:
Use this for open-ended questions that require retrieval and reasoning across datasets.
Example request body:
{
"username": "self",
"namespace": "<NAMESPACE>",
"query": "What were our top 3 entities last month?"
}
{
"response": "The top three entities last month were A, B, and C.",
"steps": [
{ "action": "retrieval", "details": {} },
{ "action": "llm_call", "details": {} }
]
}
- If reasoning is disabled, return only the response string
- If reasoning is enabled, return both the response and the steps
Step 4: Display the Results in Your App
How you display results depends on your app builder:
- For chat interfaces, render the response text directly
- For dataset queries, loop over the
data array and show rows in a table or list
- Optionally hide fields that start with
_, since these are system metadata
Step 5: Enable Dataset Discovery (Optional)
To help users explore available datasets, you can add discovery features.
List all datasets:
List datasets in a namespace:
Read raw dataset rows:
GET /data/self/{namespace}/{objectname}
Example all-in-one prompt
Some app builders support placing all of the configuration details into a single initial prompt. This approach allows you to set up an external chatbot in one step, followed by a prompt to enter your API token value when required.
Below is an example prompt you can use with Lovable to create the chatbot in a single attempt after filling in your details:
Enter your starter prompt here, for example, "Create a simple chatbox interface that connects to the <insert-dataset-name-here> and <insert-dataset-name-here> datasets within the <insert-namespace-name-here> namespace. The chatbox should perform AutoRAG queries on the data in DataLinks when queried."
Here is connection info:
DATALINKS API — CONNECTION INSTRUCTIONS
Base URL https://api.datalinks.com/api/v1
Authentication All requests require a single HTTP header: Authorization: Bearer <DL_API_KEY>
No additional headers are required.
All requests and responses use JSON.
AVAILABLE ENDPOINTS
See https://docs.datalinks.com/api-reference for all available endpoints, some examples are included below:
Structured Data Query POST /query
Use this to run explicit or ontology-based queries against a dataset.
Request body: { "username": "self", "namespace": "<NAMESPACE>", "dataset": "<OBJECT_NAME>", "query": "Ontology(<NAMESPACE>/<OBJECT_NAME>)" }
Typical response: { "data": [ { "field1": "value", "field2": 123, "field3": true } ] }
Natural Language Data Query POST /query
Use this to ask questions in plain English against a dataset.
Request body: { "username": "self", "namespace": "<NAMESPACE>", "dataset": "<OBJECT_NAME>", "naturalLanguageQuery": "Show me the latest 50 records" }
Response format is identical to structured queries: { "data": [ ...records... ] }
AutoRAG Question Answering POST /query/autorag
Use this for open-ended questions that require retrieval plus reasoning.
Request body: { "username": "self", "namespace": "<NAMESPACE>", "query": "What were our top 3 entities last month?" }
Response: { "response": "<natural language answer>", "steps": [ { "action": "retrieval", "details": {} }, { "action": "llm_call", "details": {} } ] }
Client behavior:
If reasoning is disabled, return only the response string
If reasoning is enabled, return (response, steps)
Read Dataset Rows GET /data/self/{namespace}/{objectname}
Returns raw rows for the specified dataset.
Response: { "data": [ ...records... ] }
Discover Available Datasets GET /data GET /data/{namespace}
Returns dataset listings and metadata for discovery.
METADATA NOTES
Fields starting with "_" are system-generated metadata. Client helpers may optionally filter these fields out.
SUMMARY
POST /query → structured or natural language dataset queries POST /query/autorag → AutoRAG question answering GET /data → dataset discovery GET /data/self/... → dataset row access
Summary
- Use a clear starter prompt to define app behavior
- Connect using the DataLinks API base URL and Bearer token
- Choose
/query for dataset queries and /query/autorag for reasoning-based answers
- Render responses based on your app’s UI pattern
