TANSTACK

If you're building data-heavy apps with data grids, and AI is somewhere in your stack, this conference is extremely relevant. "Beyond the Prompt" on May 19 in London One track, no keynote theater, just engineers talking about production reality ag-grid.com/campaigns/beyo…

Folks keep asking for an IRL use case for @tan_stack Start RSCs. Ok, here you go; partial page caching. CDN cache bust or TTL your pages at a component level granularity. I kept talking about it, but not showing it, so here is a deep dive video with code.

It's been another couple months since I shared progress about TanStack Table V9, but I've been working on it almost everyday with help from @riccardoperra0 this year! Here's what we mostly have done in the alpha: ✅ React Compiler compliance (harder than you'd think 🙃) ✅ State management overhaul with TanStack Store (atoms) ✅ Optionally subscribe to state slices (Less useMemo hacks needed) ✅ Tree shakable features/plugins (reduces bundle sizes by half for an average table) ✅ Large reductions in memory usage by storing row/column/cell functions on prototypes (up to 80% memory reduction on 100k virtualized rows) ✅ New helper utils (similar to TanStack Form) like createTableHook/useAppTable, tableOptions, improved column helpers, etc. ✅ New Table Devtools and better debugging options And honestly, so much more, including a docs & examples overhaul that is still underway. We're still figuring out some details with the reactivity system, especially for non-React adapters, but I think we're getting somewhat close to beta.

Find the full blog post here: tanstack.com/blog/tanstack-…

Every AI audio provider ships a different SDK. TanStack AI just unified them. Gemini Lyria, fal MiniMax, Stable Audio. Same streaming call, any model. Typed. Tree-shakeable. Hooks for React, Solid, Vue, Svelte. Try it → tanstack.com/ai Blog 👇

Skills pull in on demand. by name. at runtime. No static paths to drift. No mappings to sync. No breakage when node_modules shifts. Same config - npm, pnpm, bun, deno, workspaces, transitive. → npm i @tanstack/intent@latest

If @tanstack/intent has EVER missed a skill on you — this update is for you 🔥 We flipped the model. No more pre-declaring every mapping in your agent config. `intent install` drops one instruction block: - `intent list` to discover, - `intent load #` to fetch.

Find the blog post here: tanstack.com/blog/debug-log…

Your AI stream is a black box. 🕳️ One flag in TanStack AI and every chunk, middleware hook, and tool call prints itself for easy debugging. 🔥 No OpenTelemetry. No dashboard. No platform. chat({ ..., debug: true }) 🚀 Blog 👇

Find the blog post on this here: tanstack.com/blog/type-safe…

Type-safe provider tools land in TanStack AI 🧩 🔒 Tools are validated per model ⚡ Invalid combos fail at compile time 🧠 No more silent fallbacks The typesafety you deserve. Blog ↓

t a n s t a c k

We've been building out TanStack Store recently. TanStack Store has become foundational as an underlying package for many of the other TanStack libraries that you probably already use, but now, we're polishing it up to make it more viable for application use.

We replaced our entire smoke-test suite with 147 deterministic E2E tests powered by aimock. 🧪 Every PR now verifies every feature across 7 LLM providers in about 2 minutes. No tokens burned. No flakiness. ✅ Here's how → tanstack.com/blog/how-we-te…

Finally, @Tan_Stack Start now supports React Server Components! Start's RSCs are a truly fetchable, cacheable and composable primitive that work with your favorite tools instead of dictating your entire architecture. Oh, and one more thing... "Composite Components" 😉 🔗⬇️🧵

Solid 2.0 beta support is now available in TanStack Router, Start, and Query. If you’re experimenting with Solid 2.0 and want routing, SSR, and data fetching in a full app setup, you can start today. Read more: tanstack.com/blog/tanstack-…

TanStack AI now ships with native skills directly in the NPM module. 🔒 Versioned & trusted 🔗 Updated with source code 🧠 TanStack Intent compatible Fastest way to try them? ⬆️ Upgrade @ tanstack/ai (and friends) 📲 Tell your agent to run: npx @tanstack/intent@latest install

LLMs are bad at math. They’re bad at orchestration. But they’re really good at writing TypeScript. So we gave them a runtime. Code Mode in TanStack AI lets the model write & execute TS instead of chaining tools. 🧠 1 call instead of N ⚡ parallel execution 📉 fewer tokens ✔︎ correct results This changes how you build AI apps. Blog ↓ tanstack.com/blog/tanstack-…