Skip to main content
1

Install CLI

pip install datafrey

datafrey
A browser opens for login. Use the magic link to sign up for DataFrey quickly.
2

Connect Snowflake database

After login, the CLI guides you through configuring the Snowflake connection.
Before you connect, read about security. Check whether DataFrey is the right fit for you, how to control your sensitive information, and how we protect your data and credentials.
  1. Enter database and warehouse
  2. Run the generated SQL in Snowflake
  3. Paste the PAT and account identifier
  4. Wait for connected status
  5. Confirm indexing — required for planning; read about security vs. quality tradeoffs
Full walkthrough: Snowflake setup.
3

Install Claude Code plugin

When prompted for an AI client, pick Claude Code. The CLI installs the plugin — MCP server and /db skill in one step.
Using Cursor or another MCP client? See AI clients.
4

Run your first query

claude
Enable the MCP server.
/mcp enable plugin:datafrey:datafrey
Run your first query.
/db how many tables i have
On the first query, Claude Code opens a browser for MCP OAuth.
5

Next step

Learn how to use DataFrey

Get the most out of /db — when to use plan, how to keep the index fresh, and tips for better answers.