Changelog

Data masking (Beta)

Neon now offers a data masking feature that makes it easy to create anonymized branches for development and testing. Define masking rules through the Neon Console or API to protect sensitive data like email addresses, names, phone numbers, and other personally identifiable information. Apply these rules to create a branch with production-like data that's safe to share with your team.

This feature uses the PostgreSQL Anonymizer extension (anon) under the hood.

For more information, see Data anonymization.

Branch auto-deletion enabled by default

When creating a branch in the Neon Console, auto-delete is now enabled by default (set to 1 day). You can uncheck this option or adjust the timeframe as needed. This new default helps reduce storage costs and prevent the accumulation of unused branches.

Console only: This change only affects branches created through the Neon Console. Branches created via API or CLI are unaffected.

This change will be rolled out to all Console users in the coming days.

For more information, see Branch expiration.

More MCP Server tools

We've added several new capabilities to the Neon MCP Server:

• Search across resources - You can now search across all your Neon resources with a single query. Ask your AI assistant:

  Can you search for "production" across my Neon resources?

  The assistant will search through organizations, projects, and branches, returning structured results with direct links to the Neon Console. Use the companion fetch tool to get detailed information about any resource.

• Read-only mode - We've added read-only mode for safe operation in cloud and production environments. Enable it by adding the x-read-only: true header to your MCP configuration:

  {
    "mcpServers": {
      "Neon": {
        "url": "https://mcp.neon.tech/mcp",
        "headers": {
          "x-read-only": "true"
        }
      }
    }
  }

  When enabled, the server restricts all operations to read-only tools. Only list and describe tools are available, and SQL queries automatically run in read-only transactions, providing a safe method for querying and analyzing production databases without any risk of accidental modifications.

• Guided onboarding - The new load_resource tool provides comprehensive getting-started guidance directly through your AI assistant. Ask "Get started with Neon" or "Help me set up my first project," and the assistant will load detailed instructions covering organization setup, project configuration, connection strings, schema creation, and migrations. This works in IDEs that don't fully support MCP resources and ensures onboarding guidance is explicitly loaded when you need it.

For more information, see Neon MCP Server.

WebSocket connection reliability improvements

WebSocket connections are now more stable during long-running queries and in edge runtime environments. The serverless proxy now prevents connection timeouts during idle periods, particularly benefiting applications with long-duration queries, analytics workloads, or deployments on edge runtimes like Vercel Edge Runtime.

Reduced compute pricing across Launch, Scale, and Agent plans

We've reduced compute pricing by up to 25% across our plans:

• Launch plan: Now $0.106 per CU-hour (previously $0.14)
• Scale plan: Now $0.222 per CU-hour (previously $0.26)
• Agent plan: Now $0.106 per CU-hour (previously $0.14)

These price reductions apply immediately to all usage going forward. No action is required—you'll automatically benefit from the new lower rates on your next billing cycle.

For detailed pricing information and usage examples, see Neon plans. To learn more about these price reductions, see our blog post.

More projects on the Free plan

We're pleased to announce that the Neon Free plan now includes:

• 20 projects
• 30 projects

This means more room for side projects, prototypes, experiments, and learning new stacks.

This change applies automatically to all Free plan users—no action required. For more information about plan limits, see Neon plans.

Connect your app to Neon with a single command

We've introduced a new command that configures the Neon MCP (Model Context Protocol) Server, giving your AI assistant full context about your Neon project—including connection details, schema, and best practices. Run this command in your app's root directory:

npx neonctl@latest init

The command walks you through an interactive setup:

npx neonctl@latest init
┌  Adding Neon to your project
│
◒  Authenticating.
┌────────┬──────────────────┬────────┬────────────────┐
│ Login  │ Email            │ Name   │ Projects Limit │
├────────┼──────────────────┼────────┼────────────────┤
│ alex   │ alex@domain.com  │ Alex   │ 20             │
└────────┴──────────────────┴────────┴────────────────┘
◇  Authentication successful ✓
│
◇  Installed Neon MCP server
│
◇  Success! Neon is now ready to use with Cursor.
│
│
◇  What's next? ────────────────────────────────────────────────────────────────────────────╮
│                                                                                           │
│  Restart Cursor and ask Cursor to "Get started with Neon using MCP Resource" in the chat  │
│                                                                                           │
├───────────────────────────────────────────────────────────────────────────────────────────╯
│
└  Have feedback? Email us at feedback@neon.tech

After setup, restart Cursor and ask your AI assistant to "Get started with Neon using MCP Resource" to launch an interactive 7-step onboarding guide. The guide analyzes your codebase and walks you through selecting or creating a project, configuring connection strings, installing dependencies (@neondatabase/serverless or pg), setting up an ORM like Prisma or Drizzle, and creating your database schema—all with contextual recommendations tailored to your environment.

This feature is currently in beta for Cursor, with VS Code and Claude Code support coming soon. For more details, see our blog post.

Backup schedule API (Beta)

You can now manage automated snapshot schedules via the Neon API. Previously, backup schedules could only be configured through the Console. With the new API endpoints, you can programmatically view and update backup schedules for your branches, enabling infrastructure-as-code workflows and automated backup management.

Available endpoints:

• GET /projects/{project_id}/branches/{branch_id}/backup_schedule — View the current backup schedule for a branch
• PUT /projects/{project_id}/branches/{branch_id}/backup_schedule — Update the backup schedule configuration

This makes it easier to standardize backup policies across projects and integrate snapshot scheduling into your deployment pipelines.

For more information, see Backup & restore.

Vercel integration now supports current Neon pricing plans

Vercel integration users now have full access to Neon's current usage-based pricing plans (Free, Launch, and Scale). Previously, Vercel integration users were limited to Neon's legacy plans. Free plan Vercel users have been automatically migrated to the new Neon Free plan. This change gives you access to all the latest features and pricing options, including:

• Usage-based pricing with $5/month minimum on paid plans
• Enhanced compute options and autoscaling
• Advanced features like database branching for preview deployments
• Flexible project and branch limits

To upgrade your plan, visit the Storage tab in your Vercel Dashboard, select your database, navigate to Settings > Update Configuration, click the integration settings link, and click Change Plan.

For more information, see Neon plans.

Snapshots now in Beta with automated scheduling

The Backup & Restore page in the Neon Console is now available to all users. It combines instant point-in-time restore with snapshots to help you protect your data and recover from accidental changes, data loss, or schema issues. The snapshots feature is now in Beta with new capabilities for automated snapshot management. Use the Enhanced view toggle to access the new Backup & Restore page with snapshot capabilities, or toggle it off to return to the previous Restore page.

Snapshots (Beta):

• Scheduled snapshots — Automate snapshots with daily, weekly, or monthly schedules (available on paid plans, excluding the Agent plan)

• Flexible retention — Configure how long to keep automated snapshots before they're automatically deleted

Instant restore improvements:

• Preview data before restoring — The instant point-in-time restore feature now includes a Preview data button that lets you browse tables, run queries, and compare schemas at the selected restore point before committing to the restore operation
• Improved timezone display — The date/time picker for instant restore now shows times in your project's region timezone, and snapshot names include the region timezone for clarity

Together, these features help you maintain reliable recovery points for your production databases without manual intervention.

Postgres extension updates

We've expanded extension support for Postgres 18.

Now available on Postgres 18:

For a complete list of Postgres extensions supported by Neon, see Postgres extensions.

100x smaller Docs pages served to LLMs

We've updated the serving logic in our documentation, guides, and PostgreSQL Tutorial sites to serve raw markdown in response to web requests from user-agents we can identify as LLM Agents (like ChatGPT, Claude, Copilot, etc...)

The result is 80-100x smaller page sizes, 6x speed-up, and most importantly fewer tokens costing you money and occupying your context windows as you work with codegen agents to ship faster.

AI-ready prompts for faster setup

We've added pre-built AI prompts to our integration guides to help you get started with Neon faster in your AI-enabled code editor. Simply copy a prompt from any guide and paste it into your AI assistant (Cursor, GitHub Copilot, Claude Code, etc.) for step-by-step setup assistance tailored to your stack.

Guides with AI prompts:

• ORMs & Drivers: TypeORM, SQLAlchemy, Neon Serverless Driver, Prisma
• JavaScript/TypeScript Frameworks: Next.js, Astro, SvelteKit, Nuxt, Solid Start, Express, NestJS, Hono, RedwoodJS, Node.js
• Python Frameworks: Django, Reflex
• PHP Frameworks: Laravel, Symfony
• Ruby: Ruby on Rails
• Java Frameworks: Quarkus (JDBC), Quarkus (Reactive), Micronaut (Kotlin)
• Languages: Python, JavaScript, Go, Java, Rust, Elixir, Elixir with Ecto, .NET, .NET Entity Framework

Each prompt provides your AI assistant with the context it needs to configure dependencies, set up environment variables, establish database connections, and create working examples.

We're actively improving these prompts based on your feedback. Share your experience through the Feedback form in the Neon Console or join the conversation on Discord.

Neon Open Source Program

We've recently launched a Neon Open Source Program to support Postgres-powered open source projects. If your open source project uses Postgres and is ready to grow, we'd love to hear from you.

Visit the Neon Open Source Program page to learn more about the benefits and apply.