Netlify

Is AI killing the maker's joy, or forcing a renaissance of real craftsmanship? That's the question Eduardo Bouças, our Distinguished Engineer and Advisor to the CTO, will be debating at Craftmatters 2026 on May 29 in Porto. He's closing out the day on a roundtable titled "Craft is dead. Long live craft" alongside builders from Subvisual and Indie Campers. If you care about what AI means for the people who actually make things, this is the conversation to be in the room for. Hosted by Pixelmatters at Hilton Porto Gaia. A full day of masterclasses, roundtables, and show & tell from teams who agonize over the details. Details and tickets: https://lnkd.in/dw_uZ9h7

The web is being built by humans and agents now. Netlify runs both. Generate with AI, write it yourself, or some of each. Deploy to real production infrastructure. Scale without thinking about it. Watch how it works.

“Netlify.ai is our first website built exclusively for AI agents and not humans.” Karthik Puvvada and Eduardo Bouças on what changes when AI agents become part of the development workflow. “We’re building a platform for humans and their agents.”

When Netlify introduced deploy previews back in 2016, the idea was that every work stream should get its own isolated environment instead of teams fighting over a shared staging server. If you had 20 features in flight, you got 20 environments. That model has covered almost every part of an application for a decade now, with one stubborn exception: the database, which stayed a single shared instance. Netlify Database closes that gap by extending the same ergonomics to the data layer, and in his latest post, Distinguished Engineer Eduardo Bouças walks through the thinking behind how it was built as an AI-native primitive. A few of the design decisions worth a closer look: → Every deploy preview gets its own database branch, so developers and their agents can experiment in complete isolation without risking production data → Migrations apply automatically at the right points in the dev cycle, keeping application code and database schema in lockstep → Agents working through Agent Runners get full access to their own branch but can never reach production, and local agents like Claude Code or Cursor only ever see read-only credentials → Local development spins up a Wasm build of Postgres, which means the engine running on your laptop is the same one running in production It's a good read on what AI-native actually means when operational safety matters as much as developer velocity. https://lnkd.in/gRkQ2bFf

Staring at a blank Netlify Database wondering what to build? Last week we shared 8 prompts to get you started. Here are 10 more — all chosen because they pair real database workloads (shared state, multi-user writes, persistent across devices) with cozy design details that make a project feel finished. Paste any of these into a starter and ship before lunch: 🍳 Family recipe box. Whole family drops in favorite recipes with photos, stories, and comments on each other's. Cozy magazine layout, hand-drawn dividers, a puff of flour that drifts across the screen the first time someone adds a recipe. 🏺 Handmade ceramics shop. Browse, add to cart, place orders that hit a tidy admin view. Gallery minimalism, terracotta accent, product photos that tilt on hover like they're being picked off the shelf. 💌 Wedding RSVP. Guests confirm, pick a meal, leave a note. Live guestbook fills as people sign in. Soft florals, elegant serif, a hand-lettered "yes!" button that wiggles. 💈 Barber booking. Services with prices, a calendar where customers grab a slot, admin view of appointments. Black and cream, sharp sans-serif, a barber-pole stripe slowly rotating in the hero. 🍕 Pizzeria ordering. Customers build a pizza, orders roll into a kitchen dashboard in real time. Wood-fired vibes, hand-lettered specials board, a flame that flickers on the logo. 📚 Book swap. Members list books, request titles, arrange pickups via short messages. Warm browns, serif type, a stack of books in the corner that grows as the catalog fills. 🌽 Farmers' market. Each vendor manages their stall, shoppers place one pickup order across multiple vendors. Hand-drawn vegetable illustrations, a tiny sun that rises across the page through the day. ✈️ Collaborative travel itinerary. Add stops, vote on activities, leave comments. Vintage-postcard aesthetic, dotted travel lines on a map, a paper airplane that loops when someone adds a destination. 🏋️ CrossFit member portal. Members book class slots, watch their streak grow, check remaining credits. Coaches see the day's roster. Black with electric yellow, a satisfying barbell-clink when a class is booked. 📷 Vintage camera auction. Sellers list with starting price and end time, buyers bid live, highest bidder gets a confirmation email. Sepia tones, film-grain texture, a soft camera-shutter click on each new bid. Every one of these needs a real database. Shared state across users, persistent across devices, consistent under load. localStorage can't do it. That's why they're paired with Netlify Database — Postgres provisioned at deploy, no setup, ready when your agent is. Start here: https://lnkd.in/dZnPqvWB

The AI era is changing more than how software gets built. It’s changing who gets to build. Elad Rosenheim and Karthik Puvvada talk about why traditional seat-based pricing models start to break down when AI tools empower more people across a company to contribute directly to development. Designers, PMs, marketers, community teams, and non-traditional technical contributors can now participate in workflows that used to belong only to engineering teams. That shift changes collaboration. And it changes the economics of software teams too.

Generative UI from a single prompt. Jack 🤔 Herrington just published a walkthrough of how Code Mode and Netlify Database work together to do something interesting. Instead of giving an LLM a tool like "execute SQL," Code Mode lets the model write TypeScript that calls injected functions to query the database, run the math, and assemble a working dashboard from UI primitives. The model is doing what models are good at (writing TypeScript) instead of what they're bad at (doing math natively). The dashboard is the output, not the goal. If you're thinking about generative UI, agent tool design, or how to give an LLM real database access without handing it raw SQL, this is worth your time. Video: https://lnkd.in/gt_Cz_GP Blog: https://lnkd.in/dZnPqvWB

This Thursday at 9 AM PT, Netlify CEO Mathias Biilmann Christensen is going live with Grace Gong on the Smart Venture Podcast. The conversation is covering a lot of ground: The Jamstack origin story, and how that early bet on decoupling the front end from the back end shaped what modern web development looks like today. Where composable architecture is headed as front-end cloud infrastructure keeps maturing, and what that means for teams shipping at scale. And how AI is reshaping developer workflows right now, from the way code gets written to the way applications get deployed. Grace runs one of the most active founder interview shows on LinkedIn, and Matt has spent two decades building developer tools. This should be a good one! 👉 RSVP: https://lnkd.in/gsU56nxF

The Netlify team has been monitoring the NPM supply chain breach, expertly covered by Socket on their blog: https://lnkd.in/e-rHhHgC We have contacted a small set of affected customers. We have no evidence that Netlify itself is affected at this time.