MCP connectorsSnowflake

Snowflake MCP: connect Snowflake to Claude and ChatGPT

Ask Claude or ChatGPT how many customers renewed last quarter, what recurring revenue looks like per plan or which accounts stopped using a feature. The AI finds the right tables in Snowflake, writes the query, runs it and answers in plain words, so the question does not have to wait for someone who knows SQL.

We build the connector for your company on Snowflake’s own API. Each person signs in with their own Snowflake user, so every query runs with that person’s role, and it only reads.

Built for your company · Works in Claude, ChatGPT and Copilot

In practice

What your team can do with Snowflake in Claude and ChatGPT

Analysts

Skip the first hour of digging

Find which schema holds what, check the columns and get a first query run in one conversation. You read the SQL and refine it yourself.
Finance

Recurring revenue without a request form

Renewals, churn and revenue per plan or customer segment, asked in plain words when the board pack is due, not a week before.
Product and sales ops

See which accounts are slipping

Usage per account, feature adoption and renewals coming up, put side by side so the right customer gets a call.
Data teams

Open the warehouse without opening everything

People ask through the AI with the roles you already manage in Snowflake. You choose which groups get the connector and which warehouse the queries run on.
Examples

Ask it like this

  • “How many customers renewed last quarter, and how many did not?”

  • “What is monthly recurring revenue per plan for the last six months?”

  • “Which accounts have not logged in for 30 days and have a renewal before year end?”

  • “Which tables in the ANALYTICS database hold subscription data, and what is in each one?”

  • “How does usage at Ashford Medical this month compare with the month before?”

  • “Show the ten customers with the most support tickets per seat this year.”

The connector

What we build for Snowflake

Deployed OS has no ready Snowflake connector yet. We build it for your company when you start, on Snowflake’s own API, with the same rules as every other connector: each person’s own login, access decided by group, every call logged.

Databases and schemas
The databases and schemas your role can see, so the AI knows where to look.
Tables and views
Tables and views with their columns, types and the comments your data team has written on them.
Read-only queries
One query at a time that only reads, run on the warehouse you choose, with a limit on how many rows come back.
Your definitions
Which tables and columns stand for a customer, a renewal or recurring revenue at your company, so two people asking get the same number.
Security

Who sees what

Read-only
The connector browses and runs queries that read. Statements that would insert, update or delete are refused. We only build write actions if you ask for them.
Each person’s own Snowflake role
Everyone signs in to Snowflake as themselves through OAuth. Snowflake checks every query against that user’s role, so a table that sales cannot query in Snowflake cannot be queried through the AI either.
Your company decides who gets it
Admins choose which groups get Snowflake in Deployed OS. Start with the analysts and add finance or sales ops when you are ready.
Every call logged
Who asked, which query was run and when. Deployed OS keeps the log, and it can be exported.
Setup

Set up the Snowflake connector in three steps

01
We go through with you which questions matter most and which tables answer them, and build the connector on Snowflake’s API.
02
Your Snowflake administrator approves the connection in your account, and you choose which groups get Snowflake in Deployed OS.
03
Each person connects with their own Snowflake user. From then on they just ask.
Compared

Snowflake’s own MCP server or Deployed OS?

Snowflake has its own Snowflake-managed MCP server. Your data team creates it inside your Snowflake account and decides what it offers, and people sign in through Snowflake OAuth. If Snowflake is the only system your team wants in its AI tool and your data team wants to run the setup, it is a good place to start.

Deployed OS is for when Snowflake is one system among several. It sits next to your CRM, finance system and other tools in one company connection, admins decide which groups get each system, every call is logged in one place, and the same rules apply in Claude and ChatGPT. We also build the tools around your own definitions, so people outside the data team can ask without knowing the table names.

Their documentation: Snowflake-managed MCP server

Questions

Snowflake MCP: common questions

Yes. Snowflake offers the Snowflake-managed MCP server, which your data team creates inside your Snowflake account. Deployed OS adds Snowflake to one company connection with your other systems, with access per group, one log and tools built for your company.

No. The connector lists databases, schemas and tables and runs queries that only read. It does not insert, update or delete anything, and we only build write actions if you ask for them.

Yes. Everyone signs in as their own Snowflake user, and Snowflake checks every query against that user’s role. The AI cannot read a table the person could not query in Snowflake.

No. They ask in plain words and the AI writes the query. It shows the SQL it ran, so an analyst can check it or carry on from there.

Deployed is an EU company and fully GDPR compliant, and you choose where your data is stored, down to the city. Query results are fetched from Snowflake when someone asks for them.

It depends on the size of your team. Get in touch and we will walk you through it.

More connectorsAll connectors
Let’s talk

See it on your own tools.

Tell us which systems your company runs on. We’ll show you what your team can ask, and what IT gets to control.