When sending messages in Lovable you can choose different modes that are good for different things.

​
Agent Mode

Agent mode enables Lovable to think, plan, and take actions autonomously. It’s designed to make Lovable more capable, reliable, and independent when handling complex development tasks. Agent mode can:
  • Interpret your request
  • Explore your codebase for the necessary context
  • Read files on demand
  • Fix and refactor code as needed
  • Debug issues proactively
  • Inspect logs and network activity to identify and fix errors automatically
  • Search the web in real-time to fetch documentation, assets, or screenshots
  • Generate and edit images for your application
All without manual intervention. With deeper codebase understanding and real-time debugging, it delivers better results with fewer errors and fewer unwanted changes. These tools allow Lovable to act with autonomy and precision, reducing friction during development. Pricing Agent Mode uses usage-based pricing, which you can read more about here. The cost of a message depends on its complexity to ensure you only pay for what you actually use. Many messages cost less than 1 credit, while more complex ones may cost more. Click here to read the announcement!

​
Chat Mode

Chat Mode is your development and debugging partner. It helps you think through problems, debug issues and plan your product directly within Lovable. Every message in Chat mode deducts 1 credit.
chat-mode-edit

Activate Chat mode

To activate Chat mode, simply click “Chat” next to where you input your messages in Lovable.
Unlike Agent mode, Chat mode is conversational and can’t make changes to your code. This is perfect for asking questions, planning your project, and debugging. It is agentic, meaning it can reason across multiple steps and decide when to search files, inspect logs, query the database and more. If you ask it for help with implementing a functionality or troubleshooting it will come up with a step-by-step plan to reach your desired outcome. You’ll get prompted with an “Implement the plan”-button. Simply press it and the AI will enter Agent mode and implement the necessary code changes. Best use cases for Chat mode
  • Efficient debugging by asking the AI to investigate issues in depth before requesting a fix.
  • Planning database structures and relationships to support effective system architecture.
  • Optimal implementation of new features by identifying the minimal necessary changes required for new functionality.
  • Requesting improvement tips to enhance your application.
  • Understanding and evaluating how potential new features will impact the application’s structure and functionality.