Oossa

Replit launches Missions to automate complex team projects

Replit's new Missions lets companies set a goal and have AI agents work within a set budget to plan, build, test and deliver results.

By Published by Oossa: 1 min read

Annie Spratt · Unsplash

Replit announced a service called Missions on Nov 17, 2026. The product is aimed at teams that lack time or capacity for big tasks such as winning enterprise customers, improving product performance, or replacing legacy software. Users define the outcome they want, set a spending limit and any approval rules, and a swarm of AI agents takes over the planning, building, testing and iteration steps.

Replit used Missions on itself to build DataCat, an in‑house application‑monitoring platform. The mission ran for two days, stayed under a $3,000 budget, and produced dashboards, alerts, tracing, and a custom monitoring agent. In a test where a database bug was introduced, DataCat raised an alert in 38 seconds, beating the 60‑second target, and traced the failure to the exact query.

Missions works by coordinating many agents that can run in parallel. Replit measured the effect of agent count on a set of 13 NumPy benchmarks. A single agent achieved up to 4× speedup; a 16‑agent swarm reached 17×; and a 32‑agent swarm hit 19×. The same pattern appeared with a different optimisation mode, showing that larger swarms explore more solutions at once.

How the service is controlled

During setup, Missions asks a series of questions to define success criteria—what counts as an enterprise customer, whether paid pilots count, or what contract value is required. The mission contract also records a spending cap and which actions need human approval, such as sending proposals or changing sales materials.

Updates can be streamed in real time, sent as a daily digest, or delivered weekly, depending on the user’s preference. An independent AI judge checks each result against the agreed criteria and feeds the feedback back to the agents. If the budget runs out before the goal is met, the mission stops and reports the incomplete outcome.

Why it matters

For a small tech team, Missions offers a way to tackle projects that would normally need multiple engineers or weeks of work, while keeping costs transparent. Companies can experiment with ambitious goals—like building a custom monitoring stack—without hiring extra staff, but they still retain final decision‑making and budget control.

Was this article useful?
Share

Read next

Oossa · Newsletter

The week in AI, explained

Every Monday: the stories worth knowing, in plain language. Free, no spam.