Overview

Connections represent links to external databases and data warehouses (Snowflake, BigQuery, Redshift, SQL Server, etc.). Through connections, you can import data into Lytics, discover external schemas, and run ad-hoc queries against your data sources.

API Reference

List Connections

GET /v2/connection

Returns all connections for the account.

Get a Connection

GET /v2/connection/{id}

Returns a single connection by ID. Credential fields are encrypted in the response.

Create a Connection

POST /v2/connection

Creates a new database connection. When a connection is created, Lytics automatically triggers a background job to compute schema embeddings for AI SQL generation.

Update a Connection

PUT /v2/connection/{id}

Updates an existing connection's configuration.

Delete a Connection

DELETE /v2/connection/{id}

Soft-deletes a connection. The connection data is retained but marked as deleted.

Query a Connection

POST /v2/connection/{id}/scan

Executes a SQL query against the external data source and returns results.

Constraints

• Rate limited: 10 requests per 10 seconds
• Row limit: Default 20 rows, maximum 100 rows
• Read-only: DROP, INSERT, UPDATE, and other write operations are blocked
• Queries are executed directly against the external database

Request Body

{
  "query": "SELECT * FROM customers WHERE signup_date > '2024-01-01' LIMIT 10"
}

Discover External Schema

List Tables

GET /v2/connection/{id}/schema

Returns a list of tables available in the external data source. Results are cached in Redis for 30 minutes.

Get Table Columns

GET /v2/connection/{id}/schema/{table}

Returns column metadata (names, types) for a specific table in the external data source. Results are cached and rate-limited.

Connection Configurations

List Connection Types

GET /v2/connectionconfig

Returns all available connection type configurations (supported database types and their required fields).

Get a Connection Type

GET /v2/connectionconfig/{id}

Returns configuration details for a specific connection type by ID or slug.