Skip to content

Replit

Build a small web app from a plain-English prompt entirely in the browser, with a database and one-click deploys included.

Websites & apps Harness

Checked

Open replit.com

What you'd use it for

Turning a plain-English description into a working web app that is already live on the internet, without separately setting up hosting, a database, or a deploy pipeline.

  • Describe an app in plain English and have Replit Agent write, run, and preview the code in the same browser tab
  • Get a database and backend provisioned automatically as part of the build, with no separate setup step
  • Publish the app to a public URL with one click from the same workspace you built it in
  • Roll a Pro-tier database back to any point in the last 28 days if an agent-made change breaks something
  • Run up to 10 agents in parallel on a Pro plan for a larger build or a team working at once

The verdict

Replit's real product is that there is nothing left to configure once the agent finishes building — the same browser tab that wrote the app is already serving it, database and hosting included. That convenience is billed separately from the subscription: every agent action draws down a checkpoint-metered credit allowance, and once that runs out billing continues automatically pay-as-you-go with no spending cap unless you set one yourself. It suits someone who wants a working app today and is willing to watch a usage dashboard for it — not someone comparing the $20 or $100 sticker price against a flat-fee seat like Cursor or Claude Code, because neither of those meters usage on top of the subscription the way Replit does.

What it costs

Tier Price For Includes
Core $20/month, or $18/month billed annually One person building and shipping small apps. AI Agent access with a monthly model-credit allowance, up to 30 hours of Free-Mode chat, up to 60 Free-Mode projects, unlimited workspaces.
Pro $100/month, or $90/month billed annually Teams needing several agents running at once and reviewers involved. Everything in Core plus 10 parallel agents, up to 15 collaborators, 50 viewer invites, 28-day database rollback, premium support.
Enterprise Custom pricing Organisations with security and compliance requirements. SSO/SAML, custom seat limits, single-tenant environments, static outbound IPs, advanced privacy controls.

Read from Replit's own pricing page on 25 Sep 2026. Vendors change prices without notice.

What you need to start

  • A Replit account, signed up with email, Google, or GitHub — no credit card needed to start on Free Mode
  • A manually-set spending cap in account settings before building anything real, since checkpoint billing has no default ceiling

About ten minutes from account signup to a first working app off a single prompt, per Replit's own quickstart guide.

Your first run Paste a one-line prompt into Replit's Agent and within about ten minutes you have a working app with a live preview, already reachable at a Replit URL.

Where it breaks

The '$20 towards the most powerful models' language on the Core tier is a credit allowance, not a fixed number of requests — how far it stretches depends entirely on which model you point the agent at.

Entirely browser-based — nothing installs locally, and the app you build is hosted and deployed from the same place you built it.

First entry in this reference. There is no previous check to compare it against.

When not to use it

Not for anyone who wants a flat, predictable monthly bill: the $20 or $100 price is a starting credit allowance that agent checkpoints draw down, and Replit does not cap pay-as-you-go overage by default, so real usage can run well past the sticker price. Also not for someone who needs the app hosted somewhere other than Replit, or triggered on a schedule with nobody watching the tab.

Most AI coding tools assume you already have somewhere to run the result — a laptop, a server, a hosting account. Replit does not make that assumption. You describe the app, it writes the code, and the same browser tab that built it is already serving it to the internet.

That collapse of build-and-host into one place is the actual product. A Core subscription is not buying a smarter agent than Cursor or Claude Code — it is buying the fact that there is nothing left to configure afterward. No deploy step, no separate database to provision, no DNS to point.

What you are actually paying for

The headline price buys a monthly allowance of “credit” toward whichever underlying model the agent uses, not a fixed number of app-builds. A small internal tool costs little; an app the agent has to revise repeatedly, against a larger codebase, costs more. Pro exists mainly for the moment one agent stops being enough — ten running in parallel, a team of collaborators, and a month of point-in-time database rollbacks in case one of those agents breaks something.

Where it stops

Replit builds and hosts on your instruction, but nothing runs on a schedule or reacts to an external trigger by default. If the plan is a workflow that fires itself at 2am without anyone opening the tab, that is Make’s job, talking to whatever Replit built through its API — not Replit’s own agent loop.

Alternatives

Get Replit

Buying the tool is not the hard part.

The ebook is about what to hand over, in what order, and how you find out when it goes wrong. One payment, no subscription.

Join the daily email.

Curated AI news for small businesses and their teams. Free, daily. Sent at 7pm SGT. Unsubscribe anytime.