</> aicodedetectAI Code Humanizer

Cursor Code Detector

Paste a snippet to check whether it looks like it was written with Cursor, the AI code editor, or by another AI assistant. Free, instant, no sign-up.

0 / 8,000 characters

Three ways Cursor writes code

Cursor is a code editor built around AI, and it produces code in three different ways. Tab completion suggests the next few lines as you type; the developer accepts them one keypress at a time. The agent, opened from the chat panel, reads the project, edits several files and can run commands. Background and cloud agents do the same work on a remote machine and push a branch or open a pull request when they finish. Each leaves different evidence, so the best way to check depends on which one you suspect.

Check the commit history first

For agent work, the repository is often more telling than the code. Users on Cursor's community forum report that agent commits get a Co-authored-by: Cursor git trailer by default, and some report a Made-with: Cursor line. In the editor and the command-line agent this can be switched off under the attribution settings, and some teams strip it with a commit hook, so its absence proves nothing. Its presence is a direct statement that the agent took part. A .cursor/rules folder or a .cursorrules file in the repository shows the project uses Cursor, though not which lines it wrote.

Signals in Cursor-written code

Cursor runs models from several providers, so its code carries the general habits of AI assistants rather than a Cursor-specific fingerprint. The detector reads those habits:

  • Edits that match the file around them. Because the agent reads the project first, its code follows local naming and structure, which makes single files harder to judge than a pasted chat answer.
  • Many files changed at once. A small request that touches tests, types and docs in one go, all with the same tidy style.
  • Edge cases nobody asked for. Empty inputs, missing keys and invalid types handled in code that was meant to cover one case.
  • Comments written for a stranger. Explanations of why a step is needed, in a file the author maintains alone.

Short snippets and Tab completions give the detector very little to read, so expect a less confident result for anything under a few dozen lines.

Read more

Frequently asked questions

Can this tool prove that Cursor wrote the code?

No. Cursor does not watermark the code it generates, so any detector reads style and gives a probability with reasons. Treat a high score as a reason to look closer, not as proof.

Can it tell Cursor apart from ChatGPT, Claude or Copilot?

Not reliably, and we do not claim to. Cursor runs models from several providers, including Anthropic and OpenAI, so its output shares the habits of those assistants. A high score means the code looks AI-written in general.

Does Cursor mark its commits?

Often, but not always. Users on Cursor's forum report that agent commits get a Co-authored-by: Cursor trailer by default, and some report a Made-with: Cursor line. In the editor and the CLI this can be turned off, so a missing trailer proves nothing.

What about Tab completions?

Inline completions are accepted keystroke by keystroke and leave no trace in git. For that kind of help, reading the code is the only option, and short completions are hard for any detector to judge.

Is my code stored?

No. The snippet goes to a third-party AI model provider (Groq, with Google Gemini as a fallback) for the analysis and the result comes straight back to you. We do not save the code or write it to our logs.