sqlike: MCP server and CLI

Check your SQL before it runs, and check whether a rewrite still returns the same results.

sqlike reads a query and tells you what is wrong with it: validity errors, anti-patterns, rewrites it can apply for you, and index advice. It also compares two queries and reports whether the second one still returns the same results. There is no model anywhere in the analysis, so the same query always gets the same answer.

This repository holds the clients: an MCP server, a CLI, and the library they share. They tokenize your SQL locally, so identifiers and literals are masked before anything leaves your machine, and forward only the tokenized query. The analysis engine is not in this repo. It runs server-side and is closed.

Dialects: Postgres, MySQL, MariaDB, SQLite, SQL Server, and DuckDB. Each rule carries a verdict measured on that engine, so the severity you get is that database's behaviour and not Postgres by inheritance. DuckDB is the columnar one: it has no general-purpose secondary index, so the ten index advisors are replaced there by four columnar ones.

Why

A lot of SQL is written by an AI now, and the SQL it writes reads well more often than it runs correctly. A LEFT JOIN turns into an INNER and rows quietly disappear. A WHERE goes missing and the UPDATE hits every row. Two tables get joined on the wrong key. None of it looks wrong on the page, and none of it shows up until it has already done something.

sqlike is the check in between. It flags unsafe patterns from a catalog of 172 rules, each one verified against a real database before it ships, and it decides whether a rewrite preserves results. That second check is sound rather than complete: it certifies the rewrites it can prove, and when it cannot prove one it answers Undecided instead of guessing. Undecided never means equivalent.

Because nothing is generated, there is no retry loop, no per-token cost, and no variance between two runs on the same input. The equivalence check normalizes rather than solves, so it answers in about a millisecond where the state-of-the-art academic prover takes hundreds. That is measured head to head, including the queries where the prover wins.

Install the MCP server

Add it to any MCP client (Claude Code, Claude Desktop, Cursor, and so on):

{
  "mcpServers": {
    "sqlike": { "command": "npx", "args": ["-y", "@sqlike/mcp"] }
  }
}

Or install it through Smithery. Set SQLIKE_API_KEY if you have a key and want the higher rate limits. Without one you get the anonymous tier, which needs no signup.

Tools

analyze

Static analysis of one query: validity, anti-patterns, suggested rewrites, and schema and index advice. Returns the JSON analysis envelope.

... View full README on GitHub