Changelog

🎉 Happy New Year, everyone! 🎉

We're excited to kick off 2025 and let you know that we're shipping again! 🚢 Here's to an amazing year ahead. Let's go!

Postgres 17 for newly created Neon projects

Postgres 17 is now the default for newly created Neon projects. Neon continues to support Postgres 14, 15, and 16, if you prefer to stick with those. For Neon's Postgres version support policy, see Postgres Version Support.

Support for pg_cron

The Postgres pg_cron extension is now available on paid Neon plans upon request. This extension lets you schedule and manage periodic jobs directly in your Postgres database. To enable pg_cron, open a support ticket and include your endpoint ID and the database name where you'd like the extension enabled. Once added, you'll need to restart your compute to make it available.

Higher connection limits for autoscaling configurations

In Postgres, the max_connections setting controls the maximum number of simultaneous client connections the database server can handle. In Neon, this setting is configured based on your compute size. With Neon's autoscaling feature, max_connections is now defined by your maximum compute size rather than your minimum compute size.

This change significantly increases connection limits for autoscaling configurations. For example, consider an autoscaling configuration with these minimum and maximum compute sizes:

• Minium: 2 CU compute size — supports up to 901 connections
• Maximum: 8 CU compute size — supports up to 3604 connections

Previously, this autoscaling configuration of 2 CU ↔ 8 CU was limited to 901 connections. Now, the same configuration allows up to 3604 connections. In this case, a 4-fold increase!

PgBouncer default_pool_size now scales

Neon supports connection pooling with PgBouncer. Previously, Neon's PgBouncer configuration set the default_pool_size to a fixed value of 64, which limited Postgres connections to 64 per user/database pair, regardless of the compute size.

Now, the default_pool_size is dynamically set to 0.9 * max_connections, enabling significantly more concurrent Postgres connections per user/database pair. Note that larger compute sizes benefit from higher max_connections limits, which result in a proportionally larger default_pool_size.

For example, on an 8 CU compute with a max_connections limit of 3604, the default_pool_size increases from 64 to 3243 (0.9 × 3604).

More support for AI agents

Neon is now available as a tool for AI agents on both AgentStack and Composio.

AgentStack lets you create AI agent projects from the command line. The Neon tool allows agents to create ephemeral or long-lived Postgres instances for structured data storage. View the Neon tool here to see how an AI agent can create a Neon database in less than 500 ms, connect to the database, and run SQL DDL and DML statements.

@tool("Create Neon Project and Database")
def create_database(project_name: str) -> str:
  """
  Creates a new Neon project. (this takes less than 500ms)
  Args:
      project_name: Name of the project to create
  Returns:
      the connection URI for the new project
  """
  try:
      project = neon_client.project_create(project={"name": project_name}).project
      connection_uri = neon_client.connection_uri(
          project_id=project.id, database_name="neondb", role_name="neondb_owner"
      ).uri
      return f"Project/database created, connection URI: {connection_uri}"
  except Exception as e:
      return f"Failed to create project: {str(e)}"

Composio lets you connect 200+ tools to AI Agents — and it now supports Neon, enabling full integration between LLMs and AI agents and Neon's API. You can find the integration here.

Neon Auth.js adapter

We've introduced an Auth.js adapter for Neon, which enables storing user and session data in your Neon database. For adapter installation and setup instructions, see Neon Adapter in the Auth.js docs.

"Perplexity mode" for the Docs

We’ve added an AI-powered "perplexity mode" to the Neon Docs site, providing a conversational interface to quickly get the answers you need.

Our AI chat assistant is built on various sources including the Neon Docs, the Neon Discord Server, and API docs — and it's updated daily.

Click Ask Neon AI to try it out.