This API provides powerful search capabilities over user profile data stored in Google BigQuery. It leverages advanced Large Language Models (LLMs) to translate natural language queries into precise SQL queries, enabling flexible and intuitive data retrieval.

WebSocket-First API

The Linkd Search API now primarily uses WebSockets for real-time search capabilities. Our WebSocket API is fully documented using AsyncAPI specifications.

​
Core Functionality

  1. Natural Language Querying: Input your search criteria in plain English.
  2. LLM-Powered SQL Generation: The API uses a configured LLM (either Azure OpenAI or Cerebras) to understand your query and generate the appropriate SQL query to run against our BigQuery database.
  3. BigQuery Execution: The generated SQL is securely executed against the user profile dataset in BigQuery.
  4. Filtered Results: The API retrieves and returns the relevant user profiles based on your query.

​
API Overview

The WebSocket API endpoint provides real-time search functionality with better performance and lower latency:

  • Endpoint: /api/search/ws
  • Protocol: WebSocket (WSS)
  • Documentation: This API is documented using AsyncAPI specification

For detailed documentation on messages, examples, and implementations, see the WebSocket API section.