It was a fun going to the Microsoft Campus this summer to hack on some early ideas here. I liked how focused on security the Windows team was. Some of those fun demos will be real products unveiled at Ignite 👀

This has been one of my favorite engineering efforts at Replit, we upgraded from containers to microVMs and rebuilt the entire control plane all without users noticing. The new infrastructure is far more secure and has unlocked powerful futures like Docker support!

It was wild to see SaaStr using the Replit Agent as their Chief Marketing Officer. They showed us at our office how Replit gives them 3 new marketing ideas a day and is hooked up to all their systems so they can judge effectiveness. They say it's much more effective than the human sales and marketing folks they hired previously.

Some myths that refuse to die: 1) The poor are getting poorer 2) Unions have been the primary driver of the reduction in work hours 3) Social spending is decreasing 4) Government officials could massively increase revenue by raising taxes on billionaires

Developing in cloud sandboxes has had nice security properties and devex for those adopted but historically didn't work for everyone. Now with agents, it's really a no brainer. If you want to generate a lot of software with certain standards, you need to streamline the tools.

Travis and his team cut customer issue resolution time by 60% in 6 months. They built a triage app on Replit to help address doubling ticket volume, getting attention to urgent issues sooner.

Replit Agent switched from Temporal to Restate for three main reasons - Lower latency and cost efficiency: Restate executes durable steps inline, adding just a few milliseconds of overhead per step. This allows Replit to make thousands of operations durable without the overhead of scheduling each one as a separate activity. - Unique building blocks for stateful, concurrent agents: Virtual Objects and Durable Coroutines simplify coordinating long-lived agent sessions, parallel subagents, user steering, approvals, and cancellation, and make the architecture easier to extend. - Restate BYOC: Fully managed deployments in Replit's own cloud account keep sensitive data private and support both its multi-tenant cloud stack and dedicated enterprise offerings. With Restate, Replit can use fast, durable execution inside the agent loop, giving them a simpler, more resilient architecture. Replit now runs tens of Restate cells across regions and availability zones, and handled a major promotional traffic spike of nearly 25 000 durable actions per second per cell a few weeks after migration. Read the full story, including how Replit built its agent architecture on Restate 🧵

i forked Excalidraw and added some new features :) 🐟 draw fish and watch them swim 🏀 draw a course and drop a ball through it with @Replit Cloud Agents, you can pull in an existing repo, make your changes, and open a PR. here’s how: 1. gave the agent the Excalidraw repo and asked it to get the app running 2. described the features i wanted, tried them in the preview, and asked for tweaks 3. created my own fork on GitHub and granted the Replit GitHub App access 4. asked the agent to run the checks and open a PR against my fork here’s the PR it created ↓

Trying out the new "Repos" feature of @Replit . Pulled Deepseek harness and set it up as a project in Replit. Asked the agent to take the reins and work on a project for me. This is something on a different level 👀 Not sure what I'm gonna do with THIS one, but I just know that this is going to open up so many possibilities...

Imagine an alien species lands on Earth. They’re incredibly intelligent, especially good with software and computers, and mostly want to help humans get things done. But they’re also unpredictable. They can misunderstand your instructions, and because they’re so powerful, a small misunderstanding can cause tremendous harm. How do we integrate them into the economy and unlock their productivity while minimizing the risk? You build a habitat where they can do great work, with boundaries that contain the damage when things go wrong. That’s how we at Replit think about AI agents. Today we’re sharing our secure Agent Habitat architecture: our blueprint for integrating agents into the enterprise.

For @Replit to become a Self-Driving Company, we have been building a strong policy and network level enforcement that goes beyond traditional guardrails. Giving agents more responsibilities means policies enforced in all layers of the system to keep humans in control.

An agent investigating an issue may need to read code, inspect o11y tools, and prepare a fix. Each step raises questions about access. Who started the work? Which systems can the agent access? What can it modify? Our latest security blog explains how Replit helps teams answer those questions and enforce the answers. Read more:

What's happening inside Replit right now is going to take the world by storm. We have sales engineers doing things only the most unhinged platform engineers could have done a few months ago. We have designers and PMs making good PRs. There is no stopping us.