Connect AI tools

Use these details after connecting, when implementing a client or troubleshooting the file workflow.

Server URL

https://pageon.cc/mcp

Transport

Streamable HTTP

Header

Authorization: Bearer YOUR_PAGEON_TOKEN

{
  "mcpServers": {
    "pageon": {
      "type": "http",
      "url": "https://pageon.cc/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_PAGEON_TOKEN"
      }
    }
  }
}

Ask the AI tool to prepare a signed upload, enhance the uploaded session, download the result JSON and save its html field as a new file, then confirm cleanup. Keep the original.

Preferred flow: create_html_upload with fileName, byte size, lang and tier; upload directly, then enhance with sessionId and credential (and the same requestId for premium). The credential argument is the short-lived upload-session credential, never your account access token. Download resultUrl JSON (or join resultUrls in order as bytes before parsing JSON), save and read back html locally, then call complete_html_upload with sessionId, credential and the matching tier.

Session calls return a download URL. Standard results contain complete HTML, a suggested filename, compatibility information, and creditsUsed: 0. Standard enhancement is free, even with zero Credits, and includes basic text editing and text notes without spending Credits. For character notes (built-in characters, custom images and random animations), advanced offline tools (font, color, opacity, duplicate, sibling drag sorting, history and reset) and debranding, explicitly authorize paid export with enhance_html_premium: 1 Credit per HTML from the credential owner's balance. Also pass a UUID requestId; retries must reuse the ID and all inputs to avoid duplicate charges. Recovery is available for 7 days while the output version is unchanged. Never switch IDs after an uncertain settlement.

Files open in browse mode. Click Edit to change text and SVG text, then Save copy to download. The Runtime stays in the file and works offline without sign-in. Original fonts, images, and scripts may still need a network connection.

Files and privacy

Files use private temporary storage. Cloud input and output are deleted after local saving is confirmed; scheduled cleanup removes leftovers. Short-lived sessions include filename, size, ownership, credential hash and expiry. Call logs contain no HTML. Original scripts are preserved but never executed on the server.