DataFrey brings text-to-SQL agent to Snowflake via MCP
A data scientist built DataFrey, a Snowflake-focused MCP server with a text-to-SQL agent, CLI setup wizard, and Claude Code integration to give AI coding assistants proper database context.
Score breakdown
Developers using Claude Code for data work can now connect it directly to Snowflake with proper schema context and a planning agent, reducing the manual SQL iteration that comes from AI tools lacking live database awareness.
- 01DataFrey is a Snowflake-only MCP server with a built-in text-to-SQL agent, created by a data scientist who found Claude Code lacking DB context.
- 02A CLI wizard handles DB connection setup and Claude Code plugin installation, replacing the complex native Snowflake MCP setup process.
- 03Users can issue natural-language slash commands such as `/db write dbt model to rank leads` inside Claude Code.
DataFrey is a Snowflake-only MCP server built by a data scientist (posting as `slava_`) who found Claude Code insufficient for SQL tasks due to a lack of database context. The project aims to be a simpler alternative to Snowflake's native MCP setup, which the author describes as painful to configure. A CLI wizard handles DB connection setup and Claude Code plugin installation in one flow, and users can issue natural-language commands such as `/db write dbt model to rank leads` directly inside Claude Code.
The core differentiator is a `plan` tool that activates when a question is ambiguous or complex.
The core differentiator is a `plan` tool that activates when a question is ambiguous or complex. It kicks off a dedicated text-to-SQL agent that uses three techniques: schema-aware RAG (built during the initial DB connection step, including some sample values), subagents that actively explore the data, and multi-step planning. The author positions this as a more effective alternative to Snowflake Cortex, which they say consistently fails to identify the right tables.
On security, the author acknowledges that exposing a database over MCP is a significant risk and says considerable effort went into safer design. By default, the CLI requests SELECT permissions across all schemas — not just `information_schema` — based on the author's view that good SQL requires peeking at actual data. The service is currently free and hosted by the author with rate limiting in place; future monetization will target planning calls above the free limit. The text-to-SQL planning component will remain closed-source, though the author plans to open-source other parts of the currently monolithic deployment.
Key facts
- 01DataFrey is a Snowflake-only MCP server with a built-in text-to-SQL agent, created by a data scientist who found Claude Code lacking DB context.
- 02A CLI wizard handles DB connection setup and Claude Code plugin installation, replacing the complex native Snowflake MCP setup process.
- 03Users can issue natural-language slash commands such as `/db write dbt model to rank leads` inside Claude Code.
- 04A `plan` tool handles complex or ambiguous queries by triggering a text-to-SQL agent that uses schema RAG, data-exploring subagents, and multi-step planning.
- 05By default, the CLI requests SELECT permissions on all schemas, not just `information_schema`, based on the author's view that good SQL requires access to actual data.
- 06The service is currently free and hosted by the author with rate limits; future paid tiers will target planning calls above the limit.