23:06
It Took Me 847 Hours in Cursor to Learn These 15 Hacks
Robin Ebers
Overview
This video presents 15 hacks for using the Cursor AI coding assistant to significantly speed up development workflows. It progresses from basic settings adjustments, like enabling usage summaries and early access features, to more advanced techniques. These include mastering keyboard shortcuts for mode and model switching, optimizing plan execution with different AI models, managing multiple projects in separate windows, and customizing appearance. The summary also covers leveraging slash commands for generating rules, integrating external documentation, safely accessing environment variables, and understanding different message queuing behaviors. Finally, it delves into creating custom AI modes, utilizing Git commit history for context, and duplicating chat states for efficient parallel development, all aimed at transforming Cursor into a more powerful and personalized coding partner.
How was this?
Save this permanently with flashcards, quizzes, and AI chat
Chapters
- Enable the 'Usage Summary' in chat settings to always display your monthly AI usage, preventing unexpected overages.
- Turn on 'Early Access' in the beta settings to get new features faster, as they are generally stable and not buggy.
- Avoid the 'Nightly' build option, even if available, due to its high bugginess, making 'Early Access' the preferred choice for most users.
These foundational settings ensure you stay informed about your AI usage and gain early access to improvements, optimizing your experience and preventing potential cost surprises.
Setting the 'Usage Summary' to 'Always' in Cursor settings allows you to see your subscription usage at a glance, similar to checking your phone's data usage.
- Use Shift + Up Arrow to quickly cycle through different Cursor modes (Ask, Agent, Plan, Background Tasks).
- Employ Command + Forward Slash to open a quick menu for switching AI models (e.g., GPT-4, Claude 4.5) and executing commands.
- Combine these shortcuts to rapidly switch modes and models, such as selecting GPT-5 for planning and then Claude 4.5 for faster execution.
Efficiently navigating between modes and AI models using keyboard shortcuts dramatically reduces time spent on repetitive actions, accelerating your coding process.
Switching to 'Plan' mode with Shift+Up, then using Command+/ to select 'GPT-5' for planning, and finally switching to 'Claude 4.5' to execute the plan quickly.
- Open multiple Cursor windows using 'File > New Window' to work on different projects or compare code side-by-side.
- Customize Cursor's appearance by enabling system theme detection and selecting preferred dark and light color themes in editor settings.
- Configure terminal behavior by splitting the terminal view to manage multiple processes simultaneously without losing context.
These hacks enhance your development environment by allowing seamless multitasking across projects and personalizing the interface for better readability and efficiency.
Having one Cursor window open for your main application code and another for its associated landing page, allowing you to edit both concurrently.
- Utilize slash commands (e.g., `/`) to access powerful built-in actions like resetting context or generating rules.
- Generate custom AI rules by providing a URL to relevant documentation or articles, which Cursor then processes into a usable rule file.
- Manage and apply specific rules to agent chats by using the 'Add' command in the agent chat interface, ensuring the AI adheres to defined guidelines.
Slash commands and custom rules allow you to guide the AI's behavior, enforce coding standards, and ensure accuracy, especially when dealing with evolving technologies like new framework versions.
Generating a rule file from a blog post about Tailwind CSS version 4 changes to prevent the AI from using outdated syntax.
- Index external documentation by providing a URL in the 'Index and Docs' settings, making the entire content searchable within Cursor.
- Access and utilize your project's Git commit history via the 'Show Commits' slash command to understand past changes and inform current decisions.
- Attach summaries of previous chats to new ones using the 'Add' command to provide historical context and improve AI responses.
Integrating external knowledge bases and Git history provides the AI with a comprehensive understanding of your project and its evolution, leading to more informed and accurate assistance.
Indexing the entire Convex documentation so that the AI can answer questions about its implementation directly from the source material.
- Configure how messages are sent while the AI is streaming: 'Send' to steer the agent, 'Q' to queue for later, or 'Stop and Send' with caution.
- Create custom AI modes (e.g., 'Architect') with specific system prompts and capabilities, tailoring the AI for specialized tasks.
- Duplicate an entire chat session using the 'Duplicate Chat' option from the three-dot menu to create an identical starting point for parallel exploration or debugging.
These advanced techniques offer fine-grained control over AI interactions, enable specialized agent behavior, and facilitate efficient parallel development by cloning contextual states.
Duplicating a chat where the AI has successfully analyzed a complex bug, allowing you to experiment with different solutions without losing the original analysis.
- Understand that sensitive environment variables (e.g., in `.env.local`) are hidden by default for security.
- Use a `cursorignore` file to selectively grant Cursor access to specific files or variables, like `.env.local`, by negating the default ignore rule.
- Exercise extreme caution when granting access to sensitive files, ideally only in test environments, to prevent accidental exposure of secrets.
This hack allows the AI to utilize environment variables for contextually relevant tasks, speeding up development, but requires careful management to maintain security.
Adding `!.env.local` to a `cursorignore` file to allow Cursor to read API keys from the `.env.local` file for testing purposes.
Key takeaways
- Proactive monitoring of AI usage is crucial to avoid exceeding subscription limits.
- Leveraging keyboard shortcuts and efficient mode/model switching can drastically reduce development time.
- Enabling early access features provides faster access to new functionalities, enhancing productivity.
- Customizing Cursor's appearance and terminal layout improves the development environment's usability.
- Slash commands and custom rules empower users to guide AI behavior and enforce project standards.
- Integrating external documentation and Git history provides the AI with rich context for better assistance.
- Advanced chat management features like duplicating chats enable efficient parallel development and experimentation.
- Careful handling of sensitive information, like environment variables, is paramount even when granting AI access.
Key terms
Usage SummaryEarly AccessShift + Up ArrowCommand + Forward SlashPlan ModeSlash CommandsCursor RulesIndex DocsCursor IgnoreSteering the AgentCustom ModesDuplicate ChatContext PrimingEnvironment Variables
Test your understanding
- How can you proactively monitor your AI usage within Cursor to prevent unexpected costs?
- What are the primary keyboard shortcuts for switching between Cursor modes and AI models, and why are they important?
- How does enabling 'Early Access' in Cursor settings benefit a user, and what is the recommended alternative for stability?
- Explain the process of generating and applying custom rules in Cursor to guide AI behavior, using an example like Tailwind CSS.
- What is the purpose of the `cursorignore` file, and what precautions should be taken when using it to access sensitive environment variables?