Releases Index

Serverless Postgres

Managed Postgres with scale-to-zero and branching for modern stacks.

This week's digestOct 5 – 11, 2026

AI agents get deeper into the database as Postgres platforms open up their tooling

PlanetScale's MCP server learned to read branch metrics and org admins gained a cap on what AI clients can query, Neon opened every AI Gateway model to prepaid credits, and Prisma's v8 release candidate made migrations stop dropping data silently.

Read the digest47 releases covered
In this issue
01Agents move closer to the metalPlanetScalePlanetScaleNeonNeon3 releases02PlanetScale makes Postgres and migrations more elasticPlanetScalePlanetScale3 releases03Prisma's v8 candidate stops dropping data quietlyPrismaPrisma2 releases04Tracing, scopes, and smaller cleanupsCLICLIClient SDKClient SDKTursoTurso5 releases
EarlierSep 28Neon opens the Free plan, PlanetScale attacks import frictionSep 21Neon's scheduled functions and Supabase's one-command local setup lead a busy serverless Postgres weekAll digests →
Week of Oct 12in progress · digest Monday
TueOct 13, 20261 release
PlanetScale
PlanetScalePlanetScale1

Azure joins AWS and GCP as PlanetScale cloud provider

Microsoft Azure is now available as a cloud provider, letting users create PlanetScale Postgres and Neki databases in Azure regions with the same features as other clouds, including high availability, branching, Query Insights, automated backups, and Metal clusters. Vitess databases are not currently available on Azure.

Details
MonOct 12, 20263 releases
PlanetScaleTurso
Daily summary

PlanetScale adds per-deploy-request aggressive cutover

PlanetScale made Vitess aggressive cutover opt-in per deploy request from both the CLI and API, while Supabase's CLI and JS SDK shipped only prerelease and version-bump churn.

  • PlanetScale deploy requests can now opt into aggressive cutover individually via a flag on create and update, changeable any time before the schema change deploys, with show and queue outputting a cutover column
  • PlanetScale's refresh-schema command now waits for the schema snapshot and surfaces poll failures instead of a generic refresh timeout
  • Supabase's CLI and JS SDK releases were prerelease builds and version bumps only, with no user-facing changes
PlanetScalePlanetScale2

Aggressive cutover now opt-in per deploy request

Deploy requests can now opt into Vitess aggressive cutover individually via the CLI and API using the --aggressive-cutover flag, changeable any time before the schema change is deployed. pscale deploy-request show and queue now display an aggressive cutover column.

Details
PlanetScale CLIv0.345.0
TursoTurso1
Tursov0.8.3
Week of Oct 514 releases
DigestAI agents get deeper into the database as Postgres platforms open up their toolingRead →
FriOct 9, 20263 releases
PlanetScale
Daily summary

PlanetScale adds branch metrics to its MCP server

PlanetScale's MCP server now surfaces branch-level CPU, memory, disk, replication lag, and connection metrics to AI assistants, while its CLI gained Postgres extension and VTGate parameter commands and partially failed Vitess deploy requests can be retried for only the failed tables.

  • PlanetScale's MCP server can now report branch metrics — CPU, memory, disk, replication lag, and connections — for Vitess, Postgres, and Neki branches via planetscale_get_branch_metrics, defaulting to the last hour.
  • PlanetScale CLI added commands to toggle Postgres extensions and manage VTGate parameters, and now accepts human-readable sizes for keyspace storage limits.
  • Partially failed Vitess deploy requests can be retried for only the failed tables, preserving progress on tables that already completed, from the deploy request page or pscale deploy-request retry.
  • Supabase's client SDK typegen now excludes computed fields from select('*') and resolves no-argument rpc overloads.
PlanetScalePlanetScale3

PlanetScale Changelog

2 updates
MCP server now pulls branch metrics into your assistant

The PlanetScale MCP server can now check database metrics via the planetscale_get_branch_metrics tool, pulling CPU, memory, disk, replication lag, and connections into your AI assistant for Vitess, Postgres, and Neki branches. The default summary covers the last hour, with memory utilization including page cache reported alongside RSS and cache.

Details
Deploy requests retry only the failed tables

Vitess deploy requests now offer a retry option for partially failed migrations: instead of restarting the entire deploy, retry restarts the migration for only the failed tables while tables that already completed keep their existing progress. Available from the deploy request page and via pscale deploy-request retry .

Details
PlanetScale CLIv0.344.0
ThuOct 8, 20264 releases
NeonPlanetScalePrisma
Daily summary

Prisma makes migrations ask before dropping data; PlanetScale caps MCP access

Prisma's v8 release candidate stops migrations from silently dropping models or fields unless confirmed, while PlanetScale let organization admins cap the query access members can grant MCP clients.

  • Prisma's migration planner and db update now prompt on every model or field a plan would drop, with --delete to confirm and a new --rename that preserves the rows.
  • PlanetScale org admins can set MCP query access to none, read-only, or read-write in organization settings, applied the next time a client is authorized.
  • Prisma's v8 ORM client also gained row locks and renamed apply/scope to with/fragment.
NeonNeon1
PlanetScalePlanetScale1

Org admins can cap MCP query access per client

Organization administrators can set MCP query access in organization settings to no query access, read-only, or read and write. The limit applies the next time someone authorizes an MCP client, and already-authorized clients keep their existing access.

Details
PrismaPrisma2
WedOct 7, 20263 releases
TursoSupabase
Daily summary

Supabase SDK patches RCE as Prisma reworks contracts

Supabase's client SDK bumped Next.js to close a remote code execution hole in its realtime example while also fixing upload and retry behavior, and Prisma's 8.0 release candidates reworked SQL contract typing and toolchain install.

  • Supabase Client SDK patched a remote code execution vulnerability in its realtime-js example by bumping Next.js, and now sends Blob or File parts with their own content type in function invocations.
  • Supabase Storage uploads send multipart fields before files, and PostgREST retry logic treats TimeoutError as an abort.
  • Prisma's contract now names each column's data type (e.g. pg/text) instead of the database-printed type name, with a script and prisma db sign to upgrade existing projects.
  • Prisma fixed a Bun install failure where duplicate ArkType copies broke prisma orm init and prisma contract emit, and now reports rejected @default values with guidance on what to write.
  • Turso CLI added scopes to org-wide API tokens.
TursoTurso2
Turso CLIv1.0.34
SupabaseSupabase1
Client SDKv2.117.3
TueOct 6, 20264 releases
TursoSupabasePlanetScale
Daily summary

PlanetScale adds per-region read replicas for Postgres

PlanetScale Postgres gained independently configurable dedicated read replicas across regions, while the Supabase CLI added opt-in OTLP tracing and persistent stack log history.

  • PlanetScale Postgres now supports dedicated read replicas configurable per region, with their own cluster size, replica count, storage, and Postgres parameters, manageable via dashboard, API, CLI, or Terraform.
  • Supabase CLI runs can now be traced end-to-end with opt-in OTLP export, and stack service logs are persisted and browsable through stack log history.
TursoTurso2
Tursov0.8.2
SupabaseSupabase1
CLIv2.120.0
PlanetScalePlanetScale1

Postgres dedicated read replicas, configurable per region

PlanetScale Postgres now supports independently configurable dedicated read replicas, created in the primary region or another region from the same cloud provider with their own cluster size, replica count, storage, and Postgres parameters. They can be managed from the dashboard, API, CLI, or Terraform.

Details
Week of Sep 282 releases
FriOct 2, 20262 releases
PlanetScale
Daily summary

Neon expands free tier and adds embedding models to its AI Gateway

Neon doubled per-project storage on its free plan and began serving embedding models from its OpenAI-compatible AI Gateway, while PlanetScale added keyspace parameter management and reparent metrics and Supabase patched a Next.js RCE in its client SDK.

  • Neon Free plan projects now get 1 GB of Postgres storage each, up from 0.5 GB, applied automatically to existing projects with the 100-project limit unchanged.
  • Neon's AI Gateway now exposes OpenAI-compatible embeddings, letting you generate vectors and query them with Lakebase Search inside a single branch.
  • PlanetScale CLI can view and change Vitess keyspace VTTablet and MySQL parameters, track each change as it rolls out, and reset to defaults.
  • PlanetScale added Prometheus metrics for planned and emergency reparents, but changed planetscale_vtorc_recovery_type label values, so queries filtering on the old planned/unplanned values need updating.
  • Supabase bumped Next.js in its client SDK to patch a remote code execution vulnerability and now sends Blobs and Files with their own content type.
PlanetScalePlanetScale2

Vitess keyspace parameters editable from CLI

The new pscale keyspace parameters command lets you view and change VTTablet and MySQL parameters of a Vitess keyspace from the terminal. pscale keyspace parameters list shows current and default values, pscale keyspace parameters changes list tracks each change as it rolls out, and --reset restores a parameter to its default.

Details
PlanetScale CLIv0.343.0