Depolarize This

An MCP server that reads a polarizing social media post through the lens of depolarization: what divides, what is checkable, who hears what, and where the common ground is.

Connect

Claude Code:

claude mcp add --transport http depolarizethis https://depolarizethis.com/mcp

Any client that takes a JSON config:

{
  "mcpServers": {
    "depolarizethis": {
      "type": "http",
      "url": "https://depolarizethis.com/mcp"
    }
  }
}

Example

Call analyze_post with:

{
  "text": "Every single one of them wants to destroy this country. They have never cared about people like us.",
  "platform": "threads",
  "reply": true
}

You get back a polarization score and the techniques used (each with the exact quote), the checkable claims with a status and the kind of source that would settle each, who the post energizes and who it alienates, shared values, and a calm reply draft of at most 450 characters.

On Threads

Reply to a polarizing post with @depolarizethis and the account answers with a short, calm reply. (Coming soon.)

Limits

• The model has no web access. Claims are judged from general knowledge only and are never verified; every result says verification: "model_judgment_only".
• It is an automated read, not a ruling. Check the suggested sources before relying on it.
• Posts are processed to produce the analysis and are not stored.