Model routers are everywhere right now: a small model reads each turn and picks which model should handle it. The inconvenient truth is that the router is always less capable than the model it’s choosing for. Replit Agent lets the model decide instead. The main agent, or core loop, chooses its subagents’ tier and effort, and adjusts its own as the task unfolds. Given that freedom, Astra hands routine implementation to less costly subagents and decides for itself where its tokens are worth spending. On both DeepSWE and Terminal-Bench, Replit Agent is Pareto-efficient against Astra on its own:…
Today, we’re announcing that Replit has acquired Atta, bringing its approach to business analysis and high-quality, purpose-built charting into Replit. Our mission is to make useful intelligence accessible to everyone. For businesses, that means giving the people closest to a problem the ability to investigate it, explain what they find, and act on their expertise. Product leaders understand their customers. Marketing teams understand their channels. Sales and Revenue Operations teams understand their pipeline. But answering questions with data often requires translating that knowledge into…
You want to bring an idea to life, not compare models. But the best model for the task can change week to week, project to project, task to task. Choosing the right one adds another decision point between your idea and the thing you’re making. Starting today, Intelligent Model Routing is available to everyone on Replit. As each task evolves, we match it with the model best suited to complete it - balancing quality, speed, and cost behind the scenes.
AI models are now capable and affordable enough to make once-unreachable outcomes practical. The opportunity for everyone to pursue and collaborate on ambitious ideas, spend less time on busywork, and raise the quality and creative output across disciplines is nearly here. The possibility is here. But for most people, the reality is not. Replit users are already building million-dollar businesses, replacing legacy software and saving hundreds of thousands of dollars, and increasing their everyday potential by creating high-quality software, designs, slides, and more.
Replit enables you to build apps to create and launch apps that hold real customer data within a day. Before coding agents, a full pre-launch security review required procuring a pen test from a vendor that would cost thousands, and weeks of back-and-forth. A penetration (”pen”) test is a safe, authorized mock cyberattack that helps you fix security flaws before bad hackers find them. As we have democratized software creation, we have also sought to democratize securing it. A new generation of creators needs to ensure their apps are hardened against attackers, who look at apps built with AI…
New Admin API, Audit Logs, and Workspace Settings give organizations more insight and flexibility into how they adopt AI with Replit. Replit helps teams turn ideas into working software quickly. Most companies start with a handful of builders and a few prototypes. Then it works, and it spreads: more teams, more projects, more software running in production. That shift creates work for a specific group of people. As AI tools spread across a company: IT, procurement, and admin teams absorb the cost. They field permission requests, run access reviews, make policy decisions tool by tool, and…
The semantic layer is the foundation AI adoption is limited by trust. A user who gets burned by a confidently wrong answer will double-check the next one, eventually routing consequential work around the system entirely. Once that happens, AI remains a tool at the edges rather than infrastructure at the center… useful, but never trusted with the workflows where its value compounds. Before a company can benefit from more capable agents, those agents need a reliable way to know what the company considers true. A semantic layer tells an agent which tables are sources of truth and how they…
In the past six months, engineers at Replit have nearly tripled code output. Review times held steady. Reversions and product incidents have stayed flat. Quality metrics improved, and releases have accelerated. All the typical trade-offs you might expect have not occurred. While the code is the visible part, what's happening under the surface is much more interesting. Agents now investigate production incidents, review pull requests, answer questions, analyze business data, triage support tickets, research sales accounts, and improve the systems that power Replit Agent itself.
Most Replit Agent users start with an idea. They describe the goal in natural language — without a repo, test suite, or chosen framework — and expect the agent to turn it into a functioning app. The result might be a website, slide deck, mobile app, several connected artifacts, or something else entirely. Vibe coders are not usually checking diffs or test output. Success for Replit Agent is deceptively simple: the app should work when users click around. That changes the job of evaluation. A single score can help with a specific shipping decision, but it cannot tell us, week over week,…
The industrial space on the Lower East Side was completely transformed. Raw concrete and exposed beams were saturated with vibrant color and pulsing lights, while deep house grooves thumped through the room. Orange bean bags dotted the floor, installations glowed in every corner, and the whole place felt alive with that unmistakable opening-night rush. Artists, filmmakers, designers, musicians, founders, and engineers descended for two days. And the crowd was the act. All around the room, people were vibe coding ideas into real things you could hold, wear, and walk out the door with. At…
Every team has conventions. How you structure a project. How you handle secrets. Your design system, your testing standards, your code style. The problem: AI Agents don’t know your conventions. So you explain it again on every prompt, paste in your standards doc, or just hope someone remembered to add the context. It is one of those small frictions that compounds quietly until you are spending more time re-teaching the Agent than actually building. Today we're launching Agent Customization: a way to give Replit Agent the context it needs to work the way you or your team actually works, across…
Today, we're excited to announce two significant updates to the Replit and Databricks integration we launched in February: the user-to-machine (U2M) connector feature is now live, and the integration is now open for public preview sign-up. ICYMI: Our February announcement introduced our joint integration for enterprise teams: the ability to build applications in Replit and deploy them directly into Databricks with the inherent governance, security, and compliance controls your organization already has in place, without any extra overhead. That initial release was powered by a…
Replit already scans your projects for vulnerable dependencies, and audits your dependencies before you publish. But risk shows up earlier than that, while you are developing, the moment a malicious package gets installed. Today we're launching Package Firewall, in partnership with Socket (a software supply-chain security company) . Package Firewall blocks malicious and compromised packages from ever being installed into your app, even while you are building. This network-level security protection eliminates any window for malware to be installed into your project. It's on by default for…
Starting today, anyone can design and launch a custom Shopify storefront by chatting with the Replit Agent. Describe the store you want, and Agent generates a custom front end, creates a new Shopify store, and adds your products from the same conversation. When you're ready to sell, claim the store in Shopify, activate payments, and use Replit to deploy your storefront to a domain. From the first prompt to taking real orders is roughly ten minutes. Why we built this Replit builders already ship real software, from internal tools, to side projects, to full businesses. But selling physical…
You shipped your app. Now what? Publishing is only the beginning. Your app may look great, but if no one can find it, it stays invisible. Getting discovered on Google or recommended by AI chatbots used to mean learning SEO and GEO from scratch- writing meta tags, guessing what crawlers want, and hoping it works. Most builders skip it entirely. But done well, SEO turns a published app into one people can actually find. What's new Today, we are introducing Replit SEO Agent, which helps your apps get discovered on web search and AI search . It runs a full scan of your app, points out what's…
We believe enterprise software should move at the speed of ideas. Today, we’re collaborating with Microsoft to make it dramatically easier for enterprises to build AI-powered apps in Replit and deploy them directly into Microsoft Fabric with enterprise-grade governance and security built in. The integration helps teams go from prompt to production faster- turning governed enterprise data into internal tools, dashboards, workflows, and AI applications in a fraction of the time traditional development requires. Bringing AI-Native Development to the Microsoft Data Ecosystem Microsoft Fabric is a…
Starting today, any organization can purchase Replit Enterprise directly on our website, configure SSO and SCIM, invite their team, and start building production apps, in minutes. No demo requests, no contract negotiations, no waiting. Get Started now. Why we built this Teams evaluating Replit Enterprise kept telling us the same thing: “We have budget. We’re ready. We just want to get started quickly.” Self-serve Enterprise removes the friction between deciding to buy and actually building, with setup—from signup to a fully provisioned SSO-enabled Enterprise account—taking less time than most…
A stream of consciousness By Haya Odeh, Co-founder & Head of Design, Replit May 2026 From the moment I got married, the question started. When are you having children? Then, once I had them: How are you still doing this? Why do you even need a job? Your husband is providing. I heard versions of that my entire career. The assumption underneath it, that a woman's ambition has a natural stopping point, that motherhood is where the building ends, is something I've been quietly arguing against for over a decade. Let me try to form my thoughts around this in answers to questions I’ve heard from…
Replit is the most secure place to vibe code. Today, we're making it dramatically easier to use the Security Center to understand your security posture across all Replit projects in your business. We’re also making it possible to remediate urgent security holes, such as active critical vulnerabilities on multiple published projects, in seconds. You can reach the Security Center from the Replit Homepage, or your Settings. When you land there, the first thing we want to answer is simple: which of my projects are actually at risk right now? Start with what's urgent At the top of Security Center,…
At Replit, we want all users to be able to build secure applications. In addition to features like Security Agent and Auto Protect, you can restrict access to your apps with Private Publishing. Private Publishing is ideal for apps that should only be accessible to a specific set of users — whether that’s a personal tool, an internal app for teammates, or an early prototype shared with close collaborators. Today, we’re introducing two updates that expand secure access for Replit builders: Private Publishing is now available to Core and Starter plan users. Previously, private deployments were…
Read at the source
Your visit, your choice.
Optional Google Analytics helps us understand visits. Microsoft Clarity records masked interactions to improve the site. Optional tools stay off unless you choose them. Privacy details.