Data and AI releases

This month1,990Companies171

1,990 this month · 171 companies

Declarative Schemas 2.0 with pg-delta

Update11 days agoUnverifiedAdded Oct 16, 2026

Supabase released Declarative Schemas 2.0 using pg-delta for its CLI. Generated migrations now automatically incorporate changes to Row Level Security policies, grants, view settings, partitions, and domains.

Topics: Postgres, Developer tools

Supabase's release notes

More Supabase releases

Every Supabase release
Betaunverified

OrioleDB is now in Public Beta

OrioleDB is now in Public Beta, allowing users to create OrioleDB projects on Pro, Team, and Enterprise plans. These projects are billed the same as standard Postgres projects and support compute and disk resizing, pause and restore, read replicas, and scheduled backups.

GAunverified

Supabase Middleware 1.0

Supabase Middleware 1.0.0 is now available on npm and JSR under semantic versioning. It is a small, MIT-licensed engine for per-request logic on Web Fetch handlers that runs on Supabase Edge Functions, Vercel Functions, Cloudflare Workers, Deno, Bun, and Node 22 or newer.

Updateunverified

PostgreSQL 15.19 / 17.11 minor release

The PostgreSQL 15.19 and 17.11 minor release is rolling out, bundling five upstream security cycles and closing 44 CVEs. Four changes may require action, including potential reindexing for ltree and btree_gist indexes, and updates for pgcrypto legacy ciphers and custom operators.

Pricingunverified

Logs usage-based pricing

Supabase Logs is moving to usage-based pricing. Over 90% of projects are within the included limits, and a grace period will extend through early 2027.

Also shipped on Oct 13, 2026

Supabase in October 2026
Betaunverified

What's included:

- Three training methods: Supervised fine-tuning (SFT) , direct preference optimization (DPO) , and reinforcement learning with group relative policy optimization (GRPO) . Each method trains a low-rank adapter (LoRA) or the full model weights. Custom loss functions : Train with any objective you compute in PyTorch from the model's log-probabilities.