Overview
Build mode (previously Agent mode) is Lovable’s autonomous execution mode, designed to implement changes directly in your project. When you give Lovable a task, it takes ownership of execution end to end. It understands your intent, explores the codebase for context, applies changes across files, and resolves issues that appear during development. While it works, you can see progress through visible tasks and review the results before moving on.What Build mode is for
You can use Build mode to:- Implement features or changes once an approach is decided
- Fix bugs or errors end to end
- Refactor code across multiple files
- Apply coordinated changes across frontend, backend, and configuration
- Debug issues that arise during implementation
- Inspect logs and network activity to identify problems
- Fetch external documentation or assets when needed
- Generate or edit images and videos for use in your project
- Verify results before finishing
How to use Build mode
Build mode is active unless you switch to Plan mode. To use it, describe what you want and send your message. Clear requirements and constraints help Lovable produce better results. Be specific about what you wantTasks and execution visibility
While Lovable is working, tasks appear in the chat interface showing:- Current step being executed
- Files being modified
- Tools being used (search, web fetch, image generation)
- Progress through multi-step implementations
- Follow progress on complex builds
- Understand the steps being performed
- Stay oriented during complex changes
- Spot issues early if something seems off
Send follow-ups mid-build
Lovable working on a request doesn’t block you. Send a follow-up, a correction, or a new idea at any time to steer the work, and Lovable picks it up at its next natural stopping point without losing completed work. Some messages, like Try to fix, run as their own request after the current one. See Send follow-ups while Lovable works for details. If you still have the now-deprecated message queue, your messages keep queueing in order instead. Send now on a queued message delivers it into the running task right away, like a follow-up. The Message queue setting in Settings → Your account lets you disable the queue and switch to follow-ups. Only accounts that still have the queue see this setting.Debugging and verification
When using Build mode, Lovable does more than ensure code compiles. As part of investigation and problem solving, the agent can:- observe build errors and failures from test runs
- inspect console output and network requests when verification tools are used
- reproduce reported issues using browser testing or targeted backend calls
- verify that fixes work as expected by running appropriate checks
How long Lovable works on one message
Lovable works on one Build mode message for up to 10 hours, so a large task, like a refactor across multiple files or a long cycle of testing and fixing, is more likely to finish in a single message. This applies automatically, and there is nothing to turn on. Most messages finish well before the limit. In the final half hour, Lovable wraps up: it finishes what is in progress instead of starting new work, and summarizes what is done and what remains, so you can ask it to continue in a new message. Because a Build mode message is billed on the work Lovable does, a message that runs for hours costs more credits than a short one. Credit check-ins still apply, so a long message pauses when it crosses your check-in level and asks whether to continue, and the stop button stops it at any time. The 10-hour limit applies to Build mode messages only. To have Lovable keep working until a large task is completely finished, set that message as a goal. Credit check-ins apply to goals too.Pricing
Pricing for Build mode is usage-based. Cost depends on factors such as:- Number of files modified
- Complexity of logic changes
- Amount of codebase exploration required
- Use of tools such as verification, browser checks, web search, or image generation
Control the cost of a Build mode request
You have several controls to shape the cost of a Build mode request:- Scope each prompt narrowly. Ask for one change at a time and name the specific files or areas to touch, as shown in How to use Build mode. Broad prompts like “clean up the app” tend to explore more of the codebase and modify more files, which costs more than a scoped prompt like “add email validation to the sign-up form in @src/pages/signup.tsx”.
- Plan first for large or unfamiliar changes. Switch to Plan mode to review the intended approach, files involved, and steps before any code is changed. Every Plan mode message costs one credit, so you can trade one predictable credit for visibility into what Build mode would do next.
- Let credit check-ins catch expensive messages. When a single message crosses your check-in level, 20 credits by default, Lovable pauses with the cost so far, and you choose whether to continue, adjust the level, or have Lovable wrap up. Manage check-ins in account settings. In beta. See Build costs.
- Stop a request in progress. Click the stop button while Lovable is working to halt the current task immediately. Stopped requests are charged only for work completed up to that point, and completed changes are kept. Use the undo button to revert to the previous state if you want to discard the changes as well.
- Watch progress in the Details view. While Lovable works, click the activity card in chat to open the Details view, which shows each step, tool, and file change as it runs, so you can stop early if the task is expanding beyond what you intended.
- Check your balance before starting large tasks. Open Settings → Plans & credit usage to see your remaining credits, or click the project name in the project editor to see the credit bar. See View your credit balance.
- Be deliberate with goals. A message you set as a goal keeps Lovable working until the goal is achieved, which can use a large number of credits. Set a goal only for a large task you want finished completely.
FAQ
Does Lovable modify my project directly when in Build mode?
Does Lovable modify my project directly when in Build mode?
What happens if errors occur during implementation?
What happens if errors occur during implementation?
Can I see what Lovable is doing while it works?
Can I see what Lovable is doing while it works?
Can multiple messages be sent while working in Build mode?
Can multiple messages be sent while working in Build mode?
How do I stop a request, and what happens when I do?
How do I stop a request, and what happens when I do?
What happens if I run out of credits while Lovable is working?
What happens if I run out of credits while Lovable is working?
When should I switch to Plan mode?
When should I switch to Plan mode?
- Think through a complex change before implementation
- Compare multiple architectural approaches
- Review or edit a plan before any code is changed
- Investigate an issue and decide on a solution first
Can I reference a specific file when chatting with Lovable?
Can I reference a specific file when chatting with Lovable?
@ and select a file, or use the Reference file in chat button in the code editor. For example:Can Lovable reuse code or features from another project?
Can Lovable reuse code or features from another project?
@ mentions to reference another project directly, or simply ask Lovable to reuse an existing implementation. The agent can access relevant code, files, assets, and chat history from other projects and adapt them to your current one.Cross-project access is read-only and respects workspace permissions.