Supabase engineering blog

Loading
Set up Supabase for every Expo environment in one command
Supabase is now an integration in the EAS CLI: running eas integrations:supabase:connect creates a Supabase project, installs the SDK, and writes keys to Expo's Production, Preview, and Development environments. The --environment preview flag gives Preview builds a separate database from Production, and Supabase security and performance advisors surface findings in the Expo build dashboard. The post is largely a feature overview with limited implementation detail.
Enterprise-managed auth for MCP: Claude can now issue the identity assertion
Supabase now lets organizations use enterprise-managed auth for its MCP server without an identity provider that supports the MCP Enterprise-Managed Authorization extension, with Claude issuing the signed ID-JAG assertion instead. The post explains the verification steps Claude and Supabase perform, plus the offboarding gap and single-issuer-per-SSO-provider constraint that admins must account for.
Introducing DBArena: open, reproducible benchmarks for hosted Postgres
DBArena is a new public benchmark site comparing hosted Postgres providers on throughput, price-performance, and p95 latency at matched vCPU counts. The post details a TPC-C-derived workload, a cache-fit versus cache-exceeding test design, and median-of-three peak-concurrency reporting, and it provides reproducible commands and open tooling. It also lists the methodology's limitations, including one workload, no think time, one region per cloud, and list-price costing.
Explorer and Notebooks: investigations your team can rerun
Supabase introduces Explorer, a dashboard SQL editor that queries both the project database and its logs, and Notebooks, versioned files in the supabase folder that save queries rather than results. Notebooks can be rerun by teammates and synced via the CLI. The feature is in preview, and an AI assistant proposes notebook edits for approval.
Multigres private alpha: automatic failover for Supabase Postgres
Supabase introduces Multigres in private alpha, adding automatic failover to Supabase Postgres through a three-node cluster spread across availability zones. The post explains how Multigres treats failover as a consensus problem over Postgres synchronous replication so acknowledged writes survive promotion, and how its built-in pooler removes the need to pick between transaction and session modes. Access is limited to paid organizations, with alpha restrictions noted.
OrioleDB public beta: a Postgres storage engine for update-heavy tables
OrioleDB, an open-source Postgres storage engine, is now in public beta on Supabase. It replaces the heap with an index-organized design, an undo log, 64-bit transaction IDs, and copy-on-write checkpoints to reduce bloat and vacuum pressure on update-heavy tables. The post also reports DBArena benchmark results against heap storage and lists the beta's limitations.
Automate your Supabase backend with Make
Supabase now integrates with Make, letting automation scenarios and AI agents query databases, read logs, and deploy Edge Functions. The post walks through two connection paths: pre-built Supabase API modules for fixed workflows, and the Supabase MCP server for agents. Each MCP tool is annotated as read-only or destructive, so users can see what a scenario can touch before it runs.
Supabase Compute: run your agents inside your project
Supabase has launched Compute in private alpha, letting developers run agents, background jobs, APIs, and WebSocket services inside their Supabase project with Node, Deno, or any Dockerfile. The post covers service sizing, CLI-based deployment, and alpha limits such as no autoscaling, no GPU, no persistent disk, and restarts that require saving long-running work to Postgres.
Getting started with the Supabase .NET SDK
The post explains how to install and configure the Supabase .NET SDK v8 across ASP.NET Core, console, desktop, mobile, and Unity applications. It covers model mapping, dependency injection, connection reuse, retries, CRUD, authentication, file uploads, session persistence, and safe API-key handling.
Scope your Personal Access Tokens
Supabase now lets developers scope Personal Access Tokens to specific organizations, projects, and permissions, with expiry and a risk rating before creation. The post explains how token permissions are bounded by the user's role, how API and MCP requests are checked, and how to build least-privilege tokens for AI agents and CI pipelines.
Give your Supabase app its own MCP server
Supabase shows how to expose an app's actions as MCP tools so that each agent call runs as the signed-in user, with Postgres RLS policies controlling what the agent can reach. The post covers the OAuth middleware pipeline, a sample tool querying an RLS-protected table, and the deployment configuration needed for external agent clients.
Operate with confidence: check production with your coding agent
Supabase describes Agent prompts, which let a coding agent run scheduled read-only checks of a production project for error spikes, security findings, slow queries, and resource growth. The post covers the four prompt types, how reports are deduplicated through saved alert state, read-only MCP setup, and the query_logs tool. It also introduces Health Check Advisors and the Database Connections dashboard page.
Scale without limits: Multigres, OrioleDB, and dbarena
Supabase introduces Multigres for PostgreSQL connection pooling and automated failover, OrioleDB for undo-log storage without table bloat or routine VACUUM, and dbarena for reproducible provider benchmarks. The post explains their architecture, scaling trade-offs, availability model, transaction-ID handling, and reported performance results.
Introducing the unified Supabase .NET SDK
Supabase consolidates its .NET packages into a single repository with unified v8 versioning across GoTrue, PostgREST, Realtime, Storage, Functions, and the umbrella client. The release also adds dependency-injection support, System.Text.Json serialization, configurable retry backoff, and a migration path for existing users.
How a new branch gets its config, migrations, and functions
Explains the steps Supabase runs when a new branch is created: which Postgres, pooler, and Auth settings are copied and when restarts happen, where migrations come from and how each file is applied as one transaction, and how Edge Functions are copied and redeployed using checksums. It also describes how failed steps block dependent steps, which helps engineers diagnose slow or failed branch setup.
Supabase Middleware: composable, reusable middleware for Web Fetch handlers
Supabase introduces @supabase/middleware, a composable, type-safe pipeline for Web Fetch handlers that lets per-request logic such as CORS, JWT verification, and scoped clients be defined once and reused across runtimes and frameworks. The post explains the shared typed context model, the defineMiddleware API, and the Postgres limitation on Cloudflare Workers caused by pg's TCP requirement. It also shows how an OAuth protected-resource entry supports MCP server authorization with RLS-scoped...
Sync real-time Stripe data to Supabase with Stripe Data Pipeline
Supabase announces that Stripe Data Pipeline can now sync Stripe payment, customer, and subscription data into a Supabase Postgres project from the Stripe Dashboard. The post explains the connection model: a scoped stripe_sdp role, short-lived JWT credentials instead of a stored database password, enforced SSL, and OAuth-approved access. It is primarily a product launch, with a modest amount of security and integration detail.
How we made the Supabase Select 2026 badge
Supabase describes how it built and shipped 800 programmable conference badges based on the Pimoroni Tufty 2350. The post covers the move from a modular prototype to an integrated board, a firmware battery-drain bug, the charging and sync logistics, and a Wi-Fi failure that forced an offline-first design. It also lists community projects built on the badge.
Declarative Schemas 2.0 with pg-delta
Supabase describes Declarative Schemas 2.0, which switches the CLI's migration generation to pg-delta, an open-source Postgres diff engine that reads the system catalog and builds a dependency graph. It now covers RLS policy alter, grants, view settings, partitions, domains, and extension-managed objects, and it warns on unsupported objects. The post also explains the engine's rewrite design choice of never parsing SQL, the extension-owned object handling for pg_cron and pgmq, and the...
Ship changes to your Supabase backend with Grok Bot
Supabase is joining the SpaceXAI Partner Network so teams can request backend changes in chat, with Grok Bot handing the work to Cursor, which builds and runs it through the Supabase MCP server. The post shares telemetry on agent tool-call patterns, such as follow-up query rates after errors and migrations, and recommends a Cursor rule that runs get_advisors after schema changes to catch missing Row Level Security.