- HIGH PROBABILITY
- LOW PROBABILITY
- UNCERTAIN
- HUMAN ACTION
Short answer
How can I tell if text or images were made by AI?
You can ask your AI agent to run Winston AI tools against specific text, files, or URLs to check for synthetic patterns. This process identifies whether the content was produced by an LLM or a generative image model.
Where every detection lands.
Detection outcomes
Your AI categorizes every scan based on the probability of synthetic origin.
HIGH PROBABILITY
Synthetic content flagged
The AI identifies strong markers of LLM or generative image production. You get a clear warning that the content is likely not human-made.
LOW PROBABILITY
Human content confirmed
The scan shows no significant signs of synthetic generation. This result suggests the content was created by a person.
UNCERTAIN
Review required
The tool finds ambiguous patterns that don't clearly lean one way. You'll need to look closer at these specific cases.
HUMAN ACTION
Manual verification needed
The AI provides the raw detection data so you can make the final call. It doesn't make the final judgment for you.
The workflow
What your AI does when content arrives.
Your agent uses Winston AI to process different types of data through specialized detection tools.
Scan text strings
Your agent takes raw text you've pasted and checks it for LLM fingerprints.
detect_ai_textAnalyze uploaded files
The AI processes documents you've uploaded to see if the writing is synthetic.
detect_ai_fileCheck web links
You can point your agent to a specific URL to scan the text found on that page.
detect_ai_urlInspect images
The agent looks at image URLs to identify if the visual was generated by an AI model.
detect_ai_image
Try it
Copy these to start.
Use these prompts to trigger a scan immediately.
Starting points
Swap the bracketed text for your actual files, links, or messages.
Winston AI Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_AHEE6L9jdtzvtQVz2jFDKMSFXIU4uMwB88JwhNGI/mcpClaude Desktop
Follow the steps below to connect in seconds.
- 1In Claude Desktop, open Settings → Connectors.
- 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
- 3Click Add and start a new chat — Winston AI capabilities are ready to use.
{
"mcpServers": {
"winston-ai-mcp": {
"url": "https://edge.vinkius.com/vk_preview_AHEE6L9jdtzvtQVz2jFDKMSFXIU4uMwB88JwhNGI/mcp"
}
}
}Check this text string for AI generation: [paste text here]
Scan this file to see if it was written by an LLM: [upload file]
Use Winston AI to check if the content at this URL is synthetic: [paste URL]
Does this image look AI-generated? Check this URL: [paste image URL]
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Winston AI once, then ask.
Just link your account once. Your credentials stay encrypted, and you can immediately start asking your agent to run scans.
Connect Winston AI to your AIFAQ
How this task behaves
- 01
Can the AI change the content it detects?
No. The tools only scan and identify content. They cannot edit or rewrite the text or images.
- 02
Does the AI automatically delete flagged content?
No. The AI only reports its findings to you. You decide what to do with the information.
- 03
Can I scan a whole website at once?
No. You must provide specific URLs for the agent to scan using the detection tools.
- 04
What happens if the AI isn't sure?
The AI will report the probability it found. It is up to you to perform a manual review if the results are ambiguous.
- 05
Can it detect AI video?
No. The current tools are limited to text, files, URLs, and images.
More questions about Winston AI? The Connector page answers them. See everything the Winston AI Connector can do
Connect Winston AI to Claude, Cursor, ChatGPT & more
