Learn how billing and credits work for Replit AI features including Agent and Assistant usage, checkpoints, and managing your spend.
Replit AI features use usage-based billing to charge you based on what you build and how much you use our AI-powered tools. This ensures you only pay for the value you receive from Agent and Assistant.
All subscribers receive one-time free Agent checkpoints and Advanced Assistant edit requests.Additional monthly credits automatically apply to AI usage:
Core subscribers: per month
Teams subscribers: per user per month
These credits cover Agent, Advanced Assistant, and other Replit cloud services like deployments, storage, and databases.
Replit Agent uses effort-based pricing that scales with the complexity of your request. You pay based on the actual work Agent performs, ensuring fair pricing whether you’re making small tweaks or building complex features.
When you chat with Agent, it creates meaningful checkpoints that capture completed work on your request:
Simple requests (like bug fixes or small changes) typically cost less than the previous fixed price
Complex builds (like full features or integrations) are bundled into one checkpoint that may cost more than but reflect the total effort involved
One checkpoint per request eliminates intermediate checkpoints and reduces billing noise
Transparent pricing shows you exactly what you’re paying for each completed task
A checkpoint occurs when Agent completes work on your request and implements the requested functionality in your code.
For detailed information about what checkpoints capture and how they work, see Checkpoints and Rollbacks.To understand the cost of sample apps, check out our simple interactive demo.
For complex projects, you can enable Dynamic Intelligence features on a per-request basis:
Extended Thinking
Part of Dynamic Intelligence. Instructs Agent to analyze your request more deeply and consider additional approaches. Best for complex architectural decisions or when you need Agent to think through multiple implementation options.
High Power Mode
Part of Dynamic Intelligence. Upgrades Agent to Claude Opus for advanced problem-solving. Ideal for complex integrations, performance optimizations, or when working with unfamiliar technologies.
These Dynamic Intelligence features increase the effort Agent expends and therefore the cost, but they’re designed to save you time and deliver better solutions for complex builds.
Request specific code changes, feature implementations, and error fixes. Each edit request that modifies your code costs and counts as a billable action.
An edit request occurs when you ask Advanced Assistant to make specific changes to your code, like building a feature or fixing an error.
You can track Assistant usage the same way as Agent usage through the Assistant tab or your usage dashboard.