• SCORE
  • METRICS
  • HUMAN
  • REPORT

How to detect AI generated text with your AI

Connect GPTZero to your agent to analyze text for synthetic patterns. Once connected, your agent handles the technical scoring for you.

Ask AI about this page

Short answer

How can I tell if text was written by an AI?

Paste the text into your chat and ask your agent to run a detection. It uses GPTZero to check for specific linguistic markers and technical metrics. You'll get a breakdown of the likelihood that the content is machine-generated.

The analysis results.

What your agent finds.

The agent processes the text and delivers these specific outcomes.

SCORE

AI Probability

The agent calculates how likely the text is to be synthetic. You get a clear percentage or score to guide your decision.

METRICS

Technical Markers

Your agent identifies specific levels of perplexity and burstiness. These metrics show how predictable or varied the writing is.

HUMAN

Manual Review

If the score is ambiguous, the agent flags the text for you. You'll need to look at these sections yourself to make a final call.

REPORT

Pattern Summary

The agent highlights where the text deviates from natural human patterns. This helps you see exactly where the writing feels robotic.

The workflow

What your AI does when text arrives.

Your agent acts as the bridge between your prompt and the detection engine.

  1. Receive input

    Your agent takes the text you provide in the chat. It prepares the content for the detection process.

    detect_ai_in_text

  2. Scan for patterns

    The agent uses the tool to look for the repetitive structures common in LLM outputs.

    detect_ai_in_text

  3. Compute metrics

    The tool calculates the perplexity and burstiness of the provided text. These numbers reveal the mathematical signature of the writing.

    detect_ai_in_text

  4. Deliver results

    The agent translates the technical scores into a plain language summary. You get a direct answer about the text's origin.

    detect_ai_in_text

Try it

Copy these to start.

Use these prompts to trigger the detection immediately.

Starting points

These are opening lines. Replace the text in brackets with the content you want to check.

GPTZero Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_Y0bXy1KzXoResEbnjX7cvEDgsbjGgKLkQUAoqZai/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — GPTZero capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "gptzero-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_Y0bXy1KzXoResEbnjX7cvEDgsbjGgKLkQUAoqZai/mcp"
    }
  }
}
Copy into chat04

  • Check this email for AI patterns: [paste email text]

  • Run a detection on this blog post draft: [paste draft]

  • Analyze the perplexity of this paragraph: [paste paragraph]

  • Is this student essay likely AI-generated? [paste essay]

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect GPTZero once, then ask.

Connect the MCP to your agent with one click. Your credentials stay encrypted on our side, and you can start checking text immediately.

Connect GPTZero to your AI

FAQ

How the task behaves

  • 01

    Can the AI rewrite the text to make it look human?

    No. The tool is for detection only. It cannot edit or change your text.

  • 02

    Does it work on files like PDFs?

    No. You need to paste the text directly into the chat for the agent to analyze it.

  • 03

    Is the detection 100% accurate?

    No. Detection relies on statistical patterns. You should use the results as a guide, not absolute proof.

  • 04

    Can the AI delete my original documents?

    No. The agent only reads the text you provide to perform the check.

  • 05

    What are perplexity and burstiness?

    Perplexity measures how predictable the text is. Burstiness measures the variation in sentence structure and length.

  • More questions about GPTZero? The Connector page answers them. See everything the GPTZero Connector can do

Connect GPTZero to Claude, Cursor, ChatGPT & more