me: ok… what do I prompt next? WeWeb AI: I got you. Just press Tab ⇥

You have a senior engineer: Claude at work 24/7. Then... Why does an experienced design engineer still build in WeWeb? "𝘞𝘦𝘞𝘦𝘣 𝘰𝘧𝘧𝘦𝘳𝘴 𝘮𝘦 𝘢 𝘴𝘵𝘳𝘶𝘤𝘵𝘶𝘳𝘦𝘥 𝘱𝘭𝘢𝘤𝘦 𝘵𝘰 𝘣𝘶𝘪𝘭𝘥. 𝘐 𝘬𝘯𝘰𝘸 𝘸𝘩𝘦𝘳𝘦 𝘦𝘷𝘦𝘳𝘺𝘵𝘩𝘪𝘯𝘨 𝘪𝘴 𝘥𝘦𝘧𝘪𝘯𝘦𝘥 𝘪𝘧 𝘐 𝘯𝘦𝘦𝘥 𝘵𝘰 𝘱𝘶𝘵 𝘮𝘺 𝘩𝘢𝘯𝘥𝘴 𝘪𝘯 𝘪𝘵.” ~ @jptrinhh He shipped Carrousl in just over three weeks. It's a photo library that turns your pictures into Instagram and TikTok carousels. He builds in Claude Code, connected to WeWeb through MCP. WeWeb handles frontend, backend, auth and database. Stripe for payments. Resend for email. Cloudflare for hosting. Here's the playbook he has used to take 10+ apps from idea to production since he started working with the WeWeb MCP in Claude: 𝟭. 𝗗𝗲𝘀𝗶𝗴𝗻 𝗳𝗶𝗿𝘀𝘁. He uses Claude to pull references from Mobbin and create mockups in Paper, working out the spec before implementation. 𝟮. 𝗕𝘂𝗶𝗹𝗱 𝗶𝗻 𝗪𝗲𝗪𝗲𝗯, 𝘃𝗶𝗮 𝗠𝗖𝗣. He established a design system early, so Claude reuses existing tokens and no-code components. 𝟯. 𝗔𝗱𝗱 𝗰𝘂𝘀𝘁𝗼𝗺 𝗰𝗼𝗺𝗽𝗼𝗻𝗲𝗻𝘁𝘀 𝘄𝗵𝗲𝗿𝗲 𝗻𝗲𝗲𝗱𝗲𝗱. His image cropper, for example, lives in GitHub. He edits the code, pushes it, and syncs the component into WeWeb. 𝟰. 𝗥𝗲𝘃𝗶𝗲𝘄 𝗶𝗻 𝘁𝗵𝗲 𝗲𝗱𝗶𝘁𝗼𝗿, 𝘁𝗵𝗲𝗻 𝘁𝗲𝘀𝘁. He checks the CSS and workflow logic before testing in preview and staging. An isolated bug gets a manual fix. An issue affecting several components goes back to Claude. 𝟱. 𝗣𝘂𝗯𝗹𝗶𝘀𝗵, 𝗲𝘅𝗽𝗼𝗿𝘁 & 𝘀𝗲𝗹𝗳-𝗵𝗼𝘀𝘁. Publishing triggers an export, and Cloudflare rebuilds the app. The agent does most of the building. WeWeb gives JP an app editor to see it, fix it and decide what ships. Here’s Carrousl 👇

⏰ Monday updates. Daily digests. Monthly cleanups. Scheduled tasks are here in WeWeb 🥳 No extra automation tool needed. Here’s how to set up your first cron job 👇

cron jobs are coming to WeWeb time to make a list 👀

A picture is worth a thousand words. Or at least one very long prompt 🖼️ Have a reference in mind? Show WeWeb AI instead of trying to put it all into words. Drag an image, PDF, or text file straight into the chat 👇

@HubSpot fans, we have news 👀 You can now connect HubSpot natively to WeWeb! ✨ And yes, WeWeb AI can build with your HubSpot data too. Give it a brief and start building the dashboards, portals, and internal tools your team needs:

Now do more with @Stripe in WeWeb. A lot more 💳 Sell a subscription → send an invoice → issue a refund → build a billing dashboard... and more! Try it out and tell us what's still missing 👀 Here’s what you could do before vs. what you can do now 👇

Prompts tell AI what to do. Taste is knowing when to say NO 🚫 Now you can protect the parts of your WeWeb project AI shouldn’t change: Pages, workflows, components, tables, APIs, variables, and more 🔒 Set your AI guardrails today:

ICYMI, we launched our @mondaydotcom integration last week 🎉 So this Monday, beat the blues. Build instead. Build that client portal. Automate that team workflow. ✨ Connect monday to WeWeb → Ask AI to build on top of your boards:

👨‍💻 Vibe coding your app? Before you ship it, run this security checklist: 🔐 AUTH & ACCESS Test like a hacker. ☐ Create two test users. Copy one user's read request as cURL, replay it in Postman with the other user's record ID. Do you get their data? ☐ Replay the same request as an update. Does their record change? ☐ Replay it as a delete action. Is their record gone? ☐ As a normal user, call an admin or privileged endpoint directly. Does the backend reject you? ☐ Change fields like `role`, `owner_id`, `user_id` or `workspace_id` in a request. Can you give yourself access you shouldn’t have? ☐ Sign out completely. Replay a write request with the authorization header removed. Does the backend accept it? 🗄️ DATA & SECRETS Consider it public until you test it. ☐ Open DevTools → Network and reload every sensitive page. Do any responses contain records or fields the current user shouldn’t receive? ☐ Inspect request headers, payloads, frontend bundles and browser storage. Can you find any private API key, service credential or secret? ☐ Are you storing sensitive data you don’t actually need? 🧹 INPUT & INTEGRATIONS Never trust user input. ☐ Replay a request without using the UI. Send missing fields, extra fields, invalid values and oversized input. Does the backend reject them safely? ☐ Upload an invalid or oversized file. Does the backend enforce the same rules as the frontend? ☐ Make a third-party integration fail or timeout in staging. Does your workflow fail safely without leaking internal details? 🚦 ABUSE PROTECTION Never assume the user “won’t find out". ☐ Repeat a harmless action several times in staging. Do login, AI, email, upload and export actions eventually hit an appropriate limit? ☐ Retry an action that should only happen once. Can it accidentally send two emails, create two payments or run twice? 🔎 LOGGING & RETESTING No app is 100% secure. Keep testing. ☐ Trigger a failed login or denied request. Can you find the event in your backend logs without passwords, tokens or secrets being logged? ☐ Change a role, workflow or data model, then rerun the permission tests. Do the same security rules still hold? 💡 And one worth remembering: Hiding a button, field or page in the frontend is UX. It is NOT security. Save this before your next launch. 🔖

Hey @NotionHQ & @n8n_io lovers 👋 Look what's new today:

🥁 𝗔𝗱𝗱 𝗔𝗜 𝘁𝗼 𝘆𝗼𝘂𝗿 𝗮𝗽𝗽 𝗶𝗻 𝗷𝘂𝘀𝘁 𝗮 𝗳𝗲𝘄 𝗰𝗹𝗶𝗰𝗸𝘀 with native integrations for @OpenAI, @AnthropicAI, and @GeminiApp. 1️⃣ Choose an action: generate text, create images, or build embeddings 2️⃣ Pick the best model for your use case 3️⃣ Connect the output to your app’s data and logic And once it’s live, 𝘁𝗿𝗮𝗰𝗸 𝗿𝗲𝗾𝘂𝗲𝘀𝘁𝘀, 𝘁𝗼𝗸𝗲𝗻 𝘂𝘀𝗮𝗴𝗲, 𝗮𝗻𝗱 𝗰𝗼𝘀𝘁𝘀 in one place. Hear the full update from @msandys_ 👇

🤯 @SpotifyEng shipped Xirp, an agent-agnostic vibe coding environment. We loved the idea, and of course, the team had to try it. So we connected the agents to WeWeb through MCP and gave them a few apps to build: → KPI Dashboard → Task Management Under the hood, they share the same global design system, so every app stays on-brand as we build. Plus, a routing skill that hands off tasks cleanly: mockups to Gemini, frontend to Claude, and backend to Codex. And the end result isn’t a throwaway prototype. It’s a secure, working WeWeb app that’s still easy to understand and edit visually when a human wants to jump in. 🫵 What are you waiting for? Bring your own agents and start building with a visual editor in the loop: weweb.io/product/mcp

😲 Wow, what a UI library. Time to bring these components into WeWeb: 1. Paste the component code into Claude 2. Ask it to build the component in your WeWeb project 3. Restyle it, wire up the logic, and make it yours ✨

🧩 Our marketing team is big fans of building HTML artifacts with Claude and ChatGPT. It’s fast. It feels interactive. No context switching. The problem starts when the wider team needs to rely on it. You can share these artifacts in one click. But suddenly it’s hard to answer: • Who can see and edit what? • Can someone on the team update this next week? • Is this a one-off file? Or is it becoming an actual internal tool? With WeWeb MCP, our team is moving those HTML artifacts into WeWeb. Same speed to build. Now the whole team can use it, maintain it, and control who gets access. And this isn’t unique to our team. Across companies, non-engineers are building and shipping the tools they need themselves. It makes sense. They’re the people closest to the problem. Then IT gets pulled in to review data security and permissions. And too often, the “safe” answer is to ban citizen development altogether. That fixes the mess. It also kills the momentum. With WeWeb, teams don’t have to choose between speed and control: → Vibe code when you want to move fast. → Open the visual editor when you need visibility into workflows, data, and access rules. Here’s a glimpse of how authentication and user data are handled 👇