Live Mode

Query your database directly. No sync, no copy, no delay.

What is Live Mode?

Live mode lets rawquery query your source database directly instead of syncing data to the lakehouse first. Under the hood, DuckDB connects to your Postgres or MySQL instance and runs queries against the live tables.

This means your query results are always up-to-date - no sync lag, no stale data.

Supported Connectors

Live mode is available for database connectors that DuckDB can connect to via scanner extensions:

PostgreSQL - via postgres_scanner
MySQL - via mysql_scanner

SaaS connectors (Stripe, HubSpot, etc.) do not support live mode - they require syncing via API.

How to Enable

Go to Connections in your dashboard
Click Add Connection and select PostgreSQL or MySQL
Enter your credentials as usual
In the Query Mode selector, choose Live
Test the connection and finish setup

You can also switch an existing connection between Sync and Live mode from the connection settings.

bash

# Via the API - create a live Postgres connection
curl -X POST /api/v1/workspaces/{workspace_id}/connections \
  -H "Authorization: Bearer rq_..." \
  -H "Content-Type: application/json" \
  -d '{
    "name": "production-db",
    "query_mode": "live",
    "credentials": {
      "type": "postgres",
      "host": "db.example.com",
      "port": 5432,
      "database": "myapp",
      "user": "rawquery",
      "password": "..."
    }
  }'

# Via the API - create a live Postgres connection
curl -X POST /api/v1/workspaces/{workspace_id}/connections \
  -H "Authorization: Bearer rq_..." \
  -H "Content-Type: application/json" \
  -d '{
    "name": "production-db",
    "query_mode": "live",
    "credentials": {
      "type": "postgres",
      "host": "db.example.com",
      "port": 5432,
      "database": "myapp",
      "user": "rawquery",
      "password": "..."
    }
  }'

Sync vs Live

	Sync	Live
Data freshness	As of last sync	Real-time
Query speed	Fast (local DuckDB)	Depends on source
Source load	None at query time	Each query hits your DB
Historical data	Yes (Iceberg snapshots)	No (current state only)
Cross-source joins	Yes	Yes (DuckDB handles federation)
Network latency	None	Depends on source location
Data syncing	Yes (manual or scheduled)	Not available - queries go direct to source

How it Works

When you query a live connection, rawquery uses DuckDB's scanner extensions to connect to your database:

Your query is parsed and the referenced tables are identified
DuckDB attaches your source database in read-only mode via postgres_scanner or mysql_scanner
Tables from the source appear in a DuckDB schema matching your connection name
DuckDB pushes predicates and projections down to the source where possible, fetching only what's needed

sql

-- If your live connection is named "prod"
-- and your Postgres has a "users" table:

SELECT * FROM prod.users WHERE active = true LIMIT 100

-- If your live connection is named "prod"
-- and your Postgres has a "users" table:

SELECT * FROM prod.users WHERE active = true LIMIT 100

When to Use Each Mode

Use Sync when:

You need historical data or time travel
Query performance matters more than freshness
You want to reduce load on your production database
You're joining with SaaS sources (Stripe, HubSpot, etc.)

Use Live when:

You need always-fresh data
Your source database can handle the additional query load
You're prototyping and don't want to wait for syncs
Your tables change frequently and sync lag is unacceptable

PostgreSQL | MySQL