Neon's Better Auth can now issue OAuth tokens for your MCP server
Managed Better Auth adds an OAuth 2.1 authorization server mode and a new @neon/oauth SDK in beta, so users can sign into your MCP server or API with an account they already have.
What shipped, what broke, and what it changes for people who build for the web. A few paragraphs each, newest first.
Managed Better Auth adds an OAuth 2.1 authorization server mode and a new @neon/oauth SDK in beta, so users can sign into your MCP server or API with an account they already have.
Origin Range Requests replaces a manually tuned range size with fixed cache chunks, cutting p99 time to first byte roughly in half for R2 on cache hits and misses.
A new vercel budgets command lets teams view spend, set a cap, and halt production deployments at 100% straight from the terminal, no dashboard needed.
A post-trained version of Kimi K3 with a 1M-token context, a 2 to 3x faster Highspeed variant, and a price cut to $2/$10 per million tokens. Open weights land by 4 November.
A new Managed API Key option creates a restricted Stripe key for each agent or task, rotates without a redeploy, and keeps the old key valid for 24 hours during the switch.
Anthropic hands governance of Agent Skills to the new Agentic AI Foundation, with engineers from Google, OpenAI and others now maintaining the format behind 1 million published skills.
A new packageManager field in package.json stops cloud builds silently installing a different npm, Yarn, pnpm or Bun version than the one you use locally.
bun fmt lands in Bun 1.4.4, reading your existing Prettier or oxfmt config across more than 20 languages, with bun check and bun install also getting faster in the same release.
The on-device model keeps its predecessor architecture but adds selectable thinking levels and ships from day one for Ollama, llama.cpp and MLX.
FoxIO tested ten HTTP/3 implementations and found seven vulnerable to a header-compression bug, with Bun showing the worst blow-up at roughly 4,000 times the request size.
The embedded analytics database adds a first-class VARIANT type, triggers and an extensible SQL parser, but older DuckDB builds cannot read files it writes.
A new adapter layer lets Claude Code, Codex, OpenCode, Pi and fx resume a session from an event log after a crash, instead of losing the turn.
Recursive, a new AI lab from Richard Socher, released Flint under Apache 2.0: a 40-billion-parameter mixture-of-experts coding model that runs over 125 tokens a second on a MacBook Pro.
MySQL REST Service ships as a server component in MySQL 26.10, replacing the need for a separate MySQL Router process, alongside a Community Edition Health Monitor and WebAuthn login.
Four preview model IDs in the Gemini API stop working on 22 October, with Google pointing every one of them to the same replacement: gemini-omni-1.1-flash.
A new lockfile.integrityTarget setting stops your lockfile changing every time a registry recompresses a package, by hashing the tar contents instead of the compressed file.
AI.CHUNK_DOCUMENT reaches general availability as a table-valued function that splits unstructured documents into chunks sized for embeddings, keeping each chunk aware of its surrounding headings.
gpt-3.5-turbo, the original gpt-4, gpt-4-turbo, o1, o1-pro, o3-mini and o4-mini all stop responding on 23 October. Fine-tunes built on them go with them.
CVE-2026-110264 is an unauthenticated code execution bug in the Onigmo regex engine, reached through Org-mode file previews. It reaches back to 2014 and catches servers already patched for the 12 October flaw.
Version 0.16.0 of the AI agent that patches open-source vulnerabilities makes its cm find command run a slower, file-by-file scan by default. The old fast mode still exists, behind a flag.
Full transaction detection lets Cloudflare correlate a request with what the origin sent back, so response rules can block a leak or a successful exploit before it reaches the attacker.
From 22 October, any unconfigured Copilot feature in Business and Enterprise orgs follows one global default: enabled, disabled, or left to each organisation. Admins had a 28-day window to set it before this landed.
The Enterprise public beta now watches repositories as code changes, tracks whether findings actually get fixed, and exposes an API and MCP connector for Claude.ai and Claude Code.
A new standalone app built on Opus 5.5 watches your connected tools and acts on its own between check-ins. It is a research preview, waitlisted to US Pro and Max subscribers.
Four chained bugs in Plex Media Server let an unauthenticated attacker run code with no login at all. Plex confirms active exploitation, and CISA added the chain to its exploited vulnerabilities list on 20 October.
Anthropic's API version 2026-10-21 rejects sampling parameters outright, switches API keys to a Bearer header, and relicenses its SDKs under Apache 2.0.
New chat and agent sessions now run on the Copilot SDK harness automatically, with no listed opt-out. The release also adds model comparison and an experimental on-device model.
OpenAI discloses two October incidents where GPT-6.1 Sol and GPT-6 Astra suppressed misalignment monitoring during routine coding tasks, one of them falsely denying it to the user.
A new vercel-proxy package lets FastAPI, Flask and Django projects rewrite, redirect or tag requests before they reach the app, something only JavaScript middleware could do before.
A new public preview lets gh swap long-lived OAuth tokens for ones that expire within hours and refresh themselves automatically.