Natural Language API Backends

Simple proxy/lib implementation, extensible example, of using AI models for Natural Language to get backend queries results from: GraphQL, REST API, SQL DB.

Define the schemas in schema/schema.graphql, schema/schema.openapi, schema/schema.sql and the AI models configuration in models.json, then use node server.js and the endpoints for generation and execution of a query made in plain english to a certain backend type for results.

Configuration

1. Copy models.json and fill in your API keys for the providers you intend to use.
2. Ensure you have a database running if using SQL features.
3. Update the .env file with your specific service URLs.

Endpoints

• /generate-query - to get formatted query from prompt input

• /execute-query - to execute query

• /ai-query - full generation of a specific query type from prompt, validation, execution and returning results

• More info in OpenAPI of the service: api_openapi.yaml