Eden

Tags: release

Database migrations involve dozens of decisions: choosing a strategy, configuring traffic routing, monitoring progress, handling failures. What if you could just talk to your migration tool?

Eden v0.13.0 introduces a Model Context Protocol (MCP) server that exposes migration operations as tools for AI assistants. Claude, ChatGPT, and other MCP-compatible clients can now manage your migrations through natural language.

What is MCP?

The Model Context Protocol is an open standard for connecting AI models to external tools and data sources. Think of it as "USB-C for AI"—a universal connector that lets any AI assistant interact with any service.

MCP became the de facto standard for agentic AI integrations after major adoption in 2025. Now Eden speaks the same language.

The Tools

Our MCP server exposes four primary operations:

list_migrations — See all your migrations with filtering and pagination

get_migration — Get details about a specific migration

define_migration — Create a new migration with strategy and data movement rules

run_migration — Execute in test or live mode

Example Conversation

You: I need to migrate our user sessions from ElastiCache to Redis Cloud. Can you help?

Claude: I'll help you set up that migration. Let me check your current migrations first...

[Calls list_migrations]

You have 3 existing migrations, none affecting the session cache. I'll create a canary migration that:
1. Starts with 5% of traffic to Redis Cloud
2. Replicates data in real-time
3. Replaces conflicts (newer writes win)

[Calls define_migration]

Migration created! Want me to run a validation test before we go live?

You: Yes, test it first

Claude: Running validation...

[Calls run_migration with mode: "test"]

All checks passed. Ready to start whenever you are.

Built-in Guardrails

When AI assistants run migrations, they need guardrails to prevent runaway failures. If error rates exceed your configured threshold, migrations automatically pause and the AI assistant will ask you how to proceed.

Getting Started

Enable MCP in your Eden configuration:

yaml

mcp:
  enabled: true
  migrations:
    enabled: true
    timeout: 30s

Then connect your AI assistant via the MCP endpoint at /api/v1/endpoints/{endpoint_id}/mcp.

MCP integration guide →