• PASS
  • FAIL
  • REVIEW
  • LOG

How to moderate text content with your AI

Connect Lingyi Wanwu to your agent to automatically verify that generated text meets your safety standards.

Ask AI about this page

Short answer

How can I moderate my AI's text outputs?

Your agent uses the check_moderation tool to scan text for policy violations. You'll get a clear report on what passed and what needs your attention. This ensures your outputs stay within your safety guidelines.

The results.

Where every check lands.

The AI categorizes every piece of text it scans.

PASS

Safe content

The AI confirms the text meets all requirements. You can use this output without further checks.

FAIL

Policy violations

The AI flags text that breaks your rules. You'll see exactly which policy was triggered.

REVIEW

Human decision

Ambiguous content is sent to you. You decide if the AI's flag was a false positive or a real issue.

LOG

Audit history

Every scan is recorded. You get a full history of what was checked and the resulting status.

The workflow

What your AI does when text arrives.

The agent acts as a safety filter for your model's responses.

  1. Analyze text

    The agent takes the generated text and runs it through the safety engine.

    check_moderation

  2. Detect violations

    The tool identifies specific categories of prohibited content like hate speech or harassment.

    check_moderation

  3. Label status

    The AI marks the content as safe or unsafe based on the policy hit.

    check_moderation

  4. Flag for you

    If the text fails, the AI marks it so you can review it manually.

    check_moderation

Try it

Copy these to start.

Use these prompts to trigger a moderation check.

Starting points

Swap the bracketed text for the actual content or specific policy names you want to check.

Lingyi Wanwu Connector

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

Connector linkhttps://edge.vinkius.com/vk_preview_JkyG9ZUTy0GPyUj394Vmol76jPwT4EVBKA0NvupX/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 — Lingyi Wanwu capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "lingyi-wanwu-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_JkyG9ZUTy0GPyUj394Vmol76jPwT4EVBKA0NvupX/mcp"
    }
  }
}
Copy into chat04

  • Check this response for policy violations: [paste text here]

  • Run a moderation check on the latest output from the Slack channel.

  • Does this text from the customer support log violate any safety rules?

  • Verify if the draft in my current project meets our content guidelines.

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

Start here

Connect Lingyi Wanwu once, then ask.

Link the MCP to your client. Your login stays encrypted on our side, and you can find the full tool list on the Connector page.

Connect Lingyi Wanwu to your AI

FAQ

How this task behaves.

  • 01

    Can the AI automatically delete bad content?

    No. The AI only identifies and flags content. You must manually delete or edit anything it marks as a violation.

  • 02

    Does the AI rewrite text to make it safe?

    No. The tool only checks the text. It does not rewrite or modify the original content.

  • 03

    What happens if the AI misses a violation?

    The AI is a tool for detection, not a perfect filter. You should still review flagged content to ensure accuracy.

  • 04

    Can I use this to moderate real-time chat?

    Yes, as long as your AI client is connected to the MCP and you prompt it to check the incoming messages.

  • 05

    Does it work with any text?

    Yes, the tool can scan any text string you provide in your prompt.

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

Connect Lingyi Wanwu to Claude, Cursor, ChatGPT & more