Snowflake to Postgres sync that serves models back to your app

Point the connector at a Snowflake warehouse and role with read access, write a query or select the tables to move, map columns to the Postgres target, and schedule incremental runs on a timestamp or sequence column. An upsert on the primary key means a rerun updates rows rather than duplicating them.

Because an application cannot query a warehouse on the hot path. Models built in Snowflake (a customer health score, a recommended price, a risk flag) are only useful once they reach the operational database the product actually reads. That direction of travel is usually called reverse ETL, and it is the most common reason for this pair.

NUMBER with a scale becomes NUMERIC and integer-scaled NUMBER becomes BIGINT, VARCHAR becomes TEXT, TIMESTAMP_NTZ becomes TIMESTAMP and TIMESTAMP_TZ becomes TIMESTAMPTZ, and VARIANT or OBJECT becomes JSONB. Watch identifier case in both directions: Snowflake folds unquoted names to uppercase and PostgreSQL folds them to lowercase, so pick lowercase targets and stay consistent.

Give the source model a reliable updated-at column or a monotonic load sequence and filter on it each run. Full reloads of a warehouse table into an operational database are expensive on both sides and lock rows the application needs. Incremental upserts on the key keep the write volume proportional to what actually changed.

The Snowflake to Postgres sync from Adapters pushes warehouse tables and dbt models from Snowflake back into a Postgres application database on an incremental schedule, casting Snowflake types to Postgres and upserting on the primary key, so your app reads modeled data with a low-latency SQL query instead of hitting the warehouse on every request. Field mapping takes about a minute and no code is required.

Yes. This Snowflake to Postgres connector ships prebuilt: the field mapping is wired the moment you pick the pair, transforms are included, and the first sync can run within minutes. No code or engineering sprint required.

Pricing is flat and monthly: Starter at $49, Growth at $149, Scale at $399. Every plan includes this pair, visual field mapping, and per-record logs. There are no per-task or per-row fees, so the bill stays the same as volume grows.

Hourly on Starter, every 5 minutes on Growth, and down to every minute on Scale. Failed records retry automatically with backoff, and alerting plus a full per-record log come standard on every run.

No. Fields are auto-mapped the moment you pick the pair, and you can rewire any mapping visually before the first sync. Incremental runs use a Snowflake watermark column so only changed rows unload; NUMBER casts to Postgres NUMERIC at its declared precision, VARIANT and ARRAY land as JSONB so nested payloads survive, uppercase Snowflake identifiers fold to your Postgres naming, TIMESTAMP_NTZ and TIMESTAMP_TZ normalize to TIMESTAMPTZ in UTC, and writes upsert with ON CONFLICT on the primary key so a re-run window updates in place instead of duplicating rows.