Advantages of Cursor Compared to Copilot

After using Cursor for a while, I feel that it directly designs functions to replace some copy-paste operations during the use of Copilot, greatly improving efficiency. It also allows for more flexible instruction and context control to achieve better code generation results.
Some specific advantages include:
  • Code can be generated based on instructions, not just repository context. Instructions can also be supplemented and modified multiple times. The Tab operation is based on repository context.
  • System-level instructions can be added to constrain all code generation. For example, code style, focus areas, etc.
  • Directly issue instructions/conversations based on selected code. For instance, modifying code in place, adding comments, explaining, etc.
  • Conduct conversations based on controllable context. For example, manually selecting files or code blocks.
  • Directly apply the code blocks generated during the conversation process to the repository. Users can choose to accept or reject changes, consistent with the merge conflicts process.
  • Directly execute commands generated during the conversation, such as installing dependencies.
  • Not only can instructions be typed in the code editor to generate code, but they can also be typed in the command box to directly generate/change execution commands.
  • Switch between generation models. The costs associated with the models need to be covered by the user, controlled via API Key.
Cursor is a code editor that is deeply integrated with AI or can be said to be developed based on AI (LLM).
Cursor requires a more flexible interface control and is not designed as a VS Code extension, so it needs to be installed separately. There are unofficial extensions, but their functionality is limited.
Instructions: instructions
Docs: https://github.com/getcursor/docs

Leave a Comment