
Software ate the world, now it's learning to run it.

ceo @replit. civilizationist

Software ate the world, now it's learning to run it.

creator of React btw. you all have no idea what he’s cooking
Just used Replit Missions to take my mobile web-application skills to the next level. Literally tearing down decade old barriers that prevented us from having high quality interactions on the web. Interactions previously exclusive only to native apps. Now unleashed. This is web!

“Get me my first customers” BALLER prompt
LLMs have gotten more autonomous recently but when you give them a goal they often misunderstand you, work aimlessly, and waste your tokens. Today, we’re introducing Missions—the world’s first agentic swarm harness that can accomplish any goal you give them within set budget. Try it today with something like “get me my first 10 customers:” replit.com/blog/introducing-…

LLMs have gotten more autonomous recently but when you give them a goal they often misunderstand you, work aimlessly, and waste your tokens. Today, we’re introducing Missions—the world’s first agentic swarm harness that can accomplish any goal you give them within set budget. Try it today with something like “get me my first 10 customers:” replit.com/blog/introducing-…


How scaling agent swarms can optimize software performance.
Agent swarms are the next scaling law in AI. We scaled pretraining, then reasoning. Now we scale the number of agents working together. On 13 NumPy optimization tasks, 1 agent hit up to 4x speedups, 16 agents reached 17x, and 32 agents reached 19x. Today we're launching Replit Missions to put swarms to work.

Agent swarms are the next scaling law in AI. We scaled pretraining, then reasoning. Now we scale the number of agents working together. On 13 NumPy optimization tasks, 1 agent hit up to 4x speedups, 16 agents reached 17x, and 32 agents reached 19x. Today we're launching Replit Missions to put swarms to work.

Agent swarms are the next scaling frontier for AI. For years, we’ve scaled intelligence by training larger models and giving them more time to think. But there’s another dimension: running many agents in parallel, exploring different approaches, challenging each other, and working together on solutions that no single agent could easily discover. OpenAI has already demonstrated the potential of this approach with breakthroughs in mathematics. Until now, experimenting with agent swarms yourself required building complicated orchestration infrastructure. Starting today on Replit, spinning up a swarm of agents is just one prompt away. Excited to see what people discover when we make collective machine intelligence programmable by anyone.
LLMs have gotten more autonomous recently but when you give them a goal they often misunderstand you, work aimlessly, and waste your tokens. Today, we’re introducing Missions—the world’s first agentic swarm harness that can accomplish any goal you give them within set budget. Try it today with something like “get me my first 10 customers:” replit.com/blog/introducing-…

Didn’t think it was possible to do iOS-quality gestures on the web
Just used Replit Missions to take my mobile web-application skills to the next level. Literally tearing down decade old barriers that prevented us from having high quality interactions on the web. Interactions previously exclusive only to native apps. Now unleashed. This is web!

AI company Anthropic has paid $1.63 million to two lobbying firms tied to Secretary of State Marco Rubio and former White House chief of staff Susie Wiles. Since April 2025, Anthropic has paid $740,000 to Continental Strategy, whose DC office is run by Rubio's former Senate chief of staff, Alberto Martinez, and where Wiles' daughter Katie works. On March 9, 2026, it hired Wiles' former firm, Ballard Partners, for $890,000. Ballard's team on the account includes former Rubio State Department adviser Micah Ketchel and former deputy DNI Amaryllis Fox Kennedy, RFK Jr.'s daughter-in-law. Weeks later, Rubio's March 30 memo set off the fall of Nvidia CEO Jensen Huang, a longtime opponent of Anthropic's push to restrict chip sales to China. Huang was then turned away from meetings with Rubio three times. Continental went on to lobby Rubio's National Security Council and State Department for Anthropic. Follow: @AFpost

I've been Mission maxxing over the last couple of weeks, I've had multiple notable achievements. My favorite was adding support of all web APIs in our durable programmatic tool calling sandbox ("codemode") for agents to leverage, barely took any of my time.
LLMs have gotten more autonomous recently but when you give them a goal they often misunderstand you, work aimlessly, and waste your tokens. Today, we’re introducing Missions—the world’s first agentic swarm harness that can accomplish any goal you give them within set budget. Try it today with something like “get me my first 10 customers:” replit.com/blog/introducing-…

Replit just launched DataCat. An open-source clone of the leading observability platform. We used a swarm of agents with @browserbase to test thousands of flows in the leading tool. Then rebuilt every flow 1:1. SaaS is dead, you can replicate anything with Replit Missions.
The leading observability platform wants $83,700/yr just to monitor your app and catch issues. So Replit Missions built DataCat: github.com/replit/datacat 100% free. This isn't a prototype. Agent swarms explored thousands of workflows in the leading observability platform and rebuilt its features 1:1.

When we shared our journey about becoming a Self Driving Company, I shared a couple examples of "loops" Engineers built to complete long-desired migrations or drive translation work across the product in record time. Back then, they were a bit hard to set up. It took an engineer dedicating time. Now, we're built it as "Missions" into Replit itself. All you need to do is describe your goal and send a swarm of Agents to accomplish it! Optimize your marketing content, migrate off that legacy database, overhaul your design system. Agents work tirelessly, they just need the right visionary to send them in the right direction. That's you. replit.com/blog/introducing-…

"Get me 10 customers" is a goal. It isn't a brief an agent should execute without boundaries. Replit has introduced Missions: agent swarms that plan, build, test and iterate toward an agreed outcome, with a spending limit and approval rules. For an agency, I'd start smaller: research one target industry and prepare three demo ideas using approved public material. Define what makes an account a fit. Require human approval before outreach, publishing, discounts or product promises. A researched lead is not a paying customer. The interesting part is the Mission Agreement: agree on success, budget and what needs permission before work starts. Replit says work stops if the budget runs out before the outcome is met. Its evaluator is another AI, not proof that a human client accepts the result. Treat the launch examples as Replit's tests, not earnings promises. I haven't run a Mission. Check access and spending terms in your account before starting; even messages during an active mission count toward its budget. replit.com/blog/introducing-… Budget caveat: the setup image says the cap covers Replit AI usage, not ads, and usage may slightly exceed it. Leave headroom rather than treating it as an exact total-cost ceiling.
LLMs have gotten more autonomous recently but when you give them a goal they often misunderstand you, work aimlessly, and waste your tokens. Today, we’re introducing Missions—the world’s first agentic swarm harness that can accomplish any goal you give them within set budget. Try it today with something like “get me my first 10 customers:” replit.com/blog/introducing-…

Replit Missions can spin up a swarm of agents to replicate any type of software of any complexity.
The leading observability platform wants $83,700/yr just to monitor your app and catch issues. So Replit Missions built DataCat: github.com/replit/datacat 100% free. This isn't a prototype. Agent swarms explored thousands of workflows in the leading observability platform and rebuilt its features 1:1.
