Replit Pricing, Replit Cost and Replit Agent Pricing for Teams

Replit has a free Starter plan with no time limit, but it is built for trying the product, not for running a business on it. Starter includes a daily Agent allowance with a monthly cap, a small monthly credit for cloud services, the Lite build mode for quick targeted edits and one published app. That app goes down automatically after 30 days and carries a "Made with Replit" badge.

What Starter leaves out is most of what a team pays for. Full build, where Agent works on its own with more autonomy, requires Core. So do third-party connectors such as Google and Stripe, Replit AI Integrations, Plan Mode, converting a Design Canvas design into an app with a backend, every artifact type beyond web and mobile apps, and removing the badge. Starter also has zero collaboration seats. Replit does not offer a time-limited free trial of Core or Pro, so the practical test is a single month of Core at $20.

Replit Agent uses effort-based pricing. When Agent completes a request in Build mode it creates one checkpoint, and that checkpoint is billed by the effort the request actually took. A bug fix or a copy change costs little. A full feature or a new integration is bundled into a single checkpoint that costs more because it did more work. Replit asks for confirmation before a paid action starts, and you can hover over any checkpoint to see what it cost.

There are three ways Agent work is charged, and they stack:

• Free Mode. Core and Pro subscribers get no-cost Agent work inside an allowance that resets every five hours and has a weekly limit. Pro gets a larger allowance than Core. The pricing page describes Core as up to 30 hours of chat and up to 60 projects on Free Mode.
• Paid modes (Power and Max). For harder work you switch to a paid mode and pick a model yourself or let Replit route it with Auto. Each checkpoint is deducted from your credits at its effort-based price. Core includes $20 toward the most powerful models and Pro includes $100.
• Agent services. When Agent calls third-party models behind the scenes, for example to generate images or use a specific model inside your app, Replit bills those calls at the provider's public API rate and deducts them from the same credit balance.

Two billing behaviors catch teams out. First, Plan Mode can perform billable reasoning on the paid path even when it does not change any code, so long planning sessions are not free outside Free Mode. Second, usage data can take up to 30 minutes to reach the dashboard, so a budget set to the exact dollar will overshoot slightly on a busy afternoon. Replit publishes no fixed price per checkpoint, which is why the budget controls matter more here than on seat-priced coding tools.

Replit Pro is worth it when you have more than 5 builders, when your monthly overage on Core passes about $80, or when you need Agent to run several tasks in parallel. Below those lines Core is the better buy, because every Pro dollar above $20 is prepaid usage you may not burn.

The break-even is simple to work out. Core at $20 returns $20 of credits, so its access fee is effectively zero. Pro at $100 monthly returns $100 of credits, so it is also zero access fee, but you commit to consuming $100 of usage. A Core team spending $20 plus $80 of overage is already paying $100 a month for the same usage, and at that point Pro gives it the same credits plus 15 builders, 10 parallel tasks, two-month rollover, a 28-day database restore window and priority support at no extra cost. On annual billing Pro gets cheaper still at $90. Rollover is the understated benefit: a team that has a slow month keeps those credits for the next sprint instead of losing them, which Core never allows.

Replit does not publish an Enterprise price, but it does publish how Enterprise is bought, and it no longer requires a sales call. Any Starter, Core or Pro account can upgrade inside the product from Settings, Billing. The wizard offers two billing models: a credit commitment, which is a fixed annual credit purchase set with a slider, or pay-as-you-go, where usage bills to the card or invoice on file once you pass your credit limit. Larger custom deals still go through sales.

What the commitment buys on top of Pro: unlimited seats with no per-seat cost, SSO and SAML with SCIM provisioning, role-based access mapped to identity-provider groups, audit logs streamed to your SIEM, admin rules that can require private deployments, ban public apps, mandate security scans and pin deployment regions, an Enterprise security center with CVE detection and SBOM downloads, first-party connections to Databricks, BigQuery, Microsoft Fabric, Hex and Snowflake, and a named product advocate and field engineer. For a US company, the triggers are usually SSO and audit logging for a security review, or more than 15 people who need to build.

Replit ships the controls, but they are off until someone turns them on. Under Account, Billing you can set usage alerts at any dollar threshold and a hard budget that stops paid usage at a cap. Organizations can raise budgets in $500 increments. If you want a team to spend only what is in a pack, Replit's own guidance is to set the limit to $0.01, so nothing bills after prepaid credits are gone.

• Default to Free Mode. Everyday edits inside the five-hour allowance cost nothing. Switch to Power or Max only for work that needs it.
• Scope before you build. A specific request produces one small checkpoint. A vague one produces one large checkpoint and often a second to fix it.
• Watch database hours, not just Agent. Always-on polling is the most common hosting surprise.
• Buy tiers, not packs. Standing overage belongs in a Pro tier at 10% to 20% off, not in packs at 0% to 5%.
• Do not build what you can buy. A lot of Replit spend goes into internal tools that repeat an ops task every week: triaging tickets, drafting reports, reconciling a spreadsheet. A ready-made agent does that job at a flat monthly price, with no credits burned on building it and no database hours to host it.

That last point is where AgentMarketplace fits. The agents in the catalog above the fold run recurring work inside the tools your team already uses, from $49 a month flat, with scoped permissions and human approval on anything that writes. Keep Replit for the product you are actually building, and hand the repetitive back-office jobs to an agent that does not bill by effort.