Integration · Data warehouse
Snowflake + OrchestrAI
Catalog exported 2026-09-02 · Snowflake website
Run Snowflake SQL from chat, load stages with COPY INTO, resize warehouses, and review query history.
OrchestrAI exposes 11 Snowflake operations: 5 are low-risk (read-only or low-impact), and 6 create or modify resources and run only after you confirm the plan. 3 of them also carry a step-level approval gate.
What teams use it for
Analytics engineers ask for a query against ANALYTICS.PUBLIC.ORDERS, check the copy history on a table that a nightly load feeds, and find the long-running statement that is holding a warehouse busy so they can cancel it. Sizing a warehouse up before a heavy backfill and back down afterwards, creating a new database for a project, and loading a stage into a table are all available; query execution is medium risk because SQL can modify data, and warehouse creation, database creation, and resizing wait for confirmation. There is no operation to suspend or resume a warehouse, grant roles, or drop objects, so access control and cleanup stay in Snowsight.
Every Snowflake operation, with its risk level
| Operation | What it does | Risk | Step-level approval |
|---|---|---|---|
List Snowflake Databases |
List databases | Low risk | No |
List Snowflake Schemas |
List schemas | Low risk | No |
List Snowflake Warehouses |
List warehouses | Low risk | No |
Snowflake Copy History |
Get copy history for a table | Low risk | No |
Snowflake Query History |
Get query history | Low risk | No |
Create Snowflake Database |
Create a Snowflake database | Creates resources | Yes |
Create Snowflake Warehouse |
Create warehouse | Creates resources | Yes |
Resize Snowflake Warehouse |
Resize warehouse | Creates resources | Yes |
Snowflake Cancel Query |
Cancel running query | Creates resources | No |
Snowflake COPY INTO |
Load data from stage into table | Creates resources | No |
Snowflake Execute Query |
Execute SQL query | Creates resources | No |
Risk tiers come from the catalog: low is read-only or low-impact, medium creates resources and is reversible, high modifies existing resources, destructive may lose data. Every plan that creates or changes resources is shown with its cost estimate and waits for your confirmation. Operations marked with a step-level approval pause again on their own step. Destructive operations require a typed risk phrase.
What you connect
A Snowflake credential (stored as snowflake).
Connected-service tokens are envelope-encrypted with a per-record key wrapped by a cloud KMS.
Prompts that work
- Run SELECT COUNT(*) FROM ANALYTICS.PUBLIC.EVENTS WHERE event_date = CURRENT_DATE() on the ANALYTICS_WH warehouse
- Resize the TRANSFORM_WH warehouse to LARGE for the backfill, I will ask you to shrink it after
- Show me the copy history for RAW.STRIPE.CHARGES for the last 24 hours and tell me if any load failed
Before anything runs
Every mutation shows its plan, cost estimate, and blast radius, then waits for your confirmation. Destructive operations require a typed risk phrase. Credentials are minted per run through OIDC federation and discarded afterward; nothing you create here is invisible later, because every resource lands in the desired-state ledger where drift is detected and can be converged. Details on the security page.
Frequently asked questions
- Is running SQL in Snowflake through OrchestrAI safe?
- snowflake_execute_query is rated medium risk because a statement can insert, update, or delete rows. Reads such as listing databases, schemas, and warehouses and reading query or copy history are low risk.
- Can OrchestrAI suspend an idle Snowflake warehouse to save credits?
- Not yet, Warehouses can be created, listed, and resized, with confirmation for create and resize, but suspend and resume are not covered. Use Snowflake's auto-suspend setting for that.
- How does OrchestrAI authenticate to Snowflake?
- You add a Snowflake credential once in the connections screen. It is envelope-encrypted with a per-record key wrapped by a cloud KMS and is only decrypted inside the run that needs it.
Related integrations
Cloud services this pairs with
Try it on your own account
Connect your cloud read-only and see your resources, drift, and costs before anything runs. $5 minimum to start. Unused credits refunded in your first 14 days.
Unused credits refunded in your first 14 days.