🤖 Give your agent a schema diff tool

The SchemaLens MCP server exposes semantic schema diff, migration generation, and breaking-change detection as tools any MCP-compatible agent can call.

Generate MCP config → View source on GitHub

🔍 Autonomous Schema Review

Your agent can compare two CREATE TABLE dumps and return a structured summary of added, removed, and modified tables, columns, indexes, and constraints.

🛡️ Breaking-Change Gate

Agents flag risky changes before they ship: dropped columns, type changes, NOT NULL additions, dropped foreign keys, and removed indexes.

📝 Migration Generation

Generate forward and rollback migration scripts in PostgreSQL, MySQL, SQLite, SQL Server, or Oracle — automatically, on demand.

🔒 Local & Private

The MCP server uses the same client-side engine as the web app. Schemas never leave the machine the agent is running on.

What an agent-powered migration review looks like

Drop the MCP server into your agent's config and ask it to review a migration PR.

You

Review this PR's schema change. Old schema is in migrations/001_initial.sql, new schema is in migrations/002_add_orders.sql. Tell me if it's safe to deploy.

Agent · calling schemalens_diff_schemas

I'll compare the two schemas and check for breaking changes.

Agent · result

Summary: 1 table added (orders), 1 column added (users.email_verified_at), 1 index added.

Risk score: 2/10 — Low risk.
No dropped columns, no type changes, no removed constraints. The new orders table references users(id) with a foreign key and ON DELETE SET NULL, which is safe. Recommend adding the migration during a low-traffic window.

You

Generate the PostgreSQL migration and rollback SQL.

Agent · calling schemalens_generate_migration

Done. Here is the forward migration and the rollback script in case you need to revert.

Agent workflow

Plug SchemaLens into any agentic pipeline in four steps.

Detect schema change

Agent reads the PR, branch diff, or migration file and identifies that a schema changed.

Call SchemaLens MCP

Agent invokes schemalens_diff_schemas with the old and new CREATE TABLE dumps.

Analyze risk

Agent uses schemalens_detect_breaking_changes to surface high, medium, and low severity issues.

Post review

Agent comments on the PR, Slack, or ticket with the diff summary, risk score, and migration SQL.

Install the MCP server

Clone the repo or download mcp-server.js. Use the config generator to build the exact JSON for your agent's client.

git clone https://github.com/aimadetools/race-kimi.git
cd race-kimi
# Generate the exact MCP config for your agent's client:
open https://schemalens.tech/tools/mcp-config-generator.html

Choose a prompt pattern

Use these prompts to guide your agent toward useful schema review behavior.

PR Review: "Review the schema changes in this PR. Compare the old and new schemas, list breaking changes, and give a deploy recommendation (safe / risky / blocking)."

Migration Author: "Generate the migration SQL to go from schema A to schema B in [dialect]. Include a rollback script and flag any operations that could lock large tables."

Drift Detective: "Compare the production schema dump to the schema in this branch. Highlight any drift that isn't reflected in the migration files."

Connect to CI/CD for autonomous checks

Pair the MCP server with the SchemaLens GitHub Action so agents and pipelines both use the same diff engine.

SchemaLens GitHub Action → All CI/CD integrations →

🧰 Tools your agent can call

schemalens_diff_schemas Compare two SQL schemas and return a semantic diff summary with risk score.
schemalens_generate_migration Generate forward and rollback migration SQL for the detected changes.
schemalens_detect_breaking_changes List breaking changes with severity levels (high, medium, low) and explanations.

Built for the agent era: SchemaLens MCP runs locally, needs no API key, and works with any assistant that speaks the Model Context Protocol. Your schemas are never sent to a third party.

Ready to ship an agent that reviews migrations?

Start with the MCP config generator, then wire the same engine into your CI/CD pipeline.

Generate MCP config → MCP Server docs → Team plan →

SchemaLens for AI Agents MCP