Automate HTML Publishing from AI Tools
AI tools generate great HTML — interactive charts, dashboards, reports, landing pages — but getting that HTML from a chat window to a live URL has always required manual steps. RunThis.page eliminates that with two integration options: MCP for AI desktop apps and a REST API for scripts and pipelines.
Both methods require a Pro ($30/mo) or Business ($50/mo) plan with an API key.
Option 1: MCP — direct AI integration
MCP (Model Context Protocol) lets AI desktop apps call external tools directly. With the RunThis.page MCP server, Claude Desktop can publish pages without you ever leaving the conversation.
How it works:
- You ask Claude to build an HTML page
- Claude generates the HTML
- Claude calls
publish_pageto upload it to RunThis.page - You get a live URL back in the conversation
Full MCP setup guide for Claude Desktop →
Option 2: REST API — scripted workflows
The REST API works from any language or platform. Use it in CI/CD pipelines, cron jobs, or custom AI agent frameworks.
curl -X POST https://runthis.page/api/v1/pages \
-H "Authorization: Bearer rtp_your_key_here" \
-H "Content-Type: text/html" \
-H "X-Page-Title: Weekly Report" \
-d @report.html
When to use which
| MCP Server | REST API | |
|---|---|---|
| Best for | Interactive AI conversations | Scripts, CI/CD, batch jobs |
| Setup | Add to Claude Desktop config | Use API key in any HTTP client |
| Languages | Any MCP-compatible AI app | Any language with HTTP support |
| Error handling | AI reads errors and recovers | Structured JSON errors in code |
| Human involvement | Conversational (approve in chat) | Fully automated (no human needed) |
Example workflows
- Daily reports — a cron job generates an HTML report from your database, publishes it via the API, and emails the link to stakeholders
- AI data analyst — ask Claude to analyze a CSV, generate an interactive chart, and publish it to share with your team
- CI/CD docs — your build pipeline generates HTML documentation on every push and publishes it to a stable URL
- Client deliverables — ask Claude to build a presentation page, review it, then share the link with your client
- Custom AI agents — build an agent that creates and publishes HTML as part of a larger workflow using the REST API
Structured error handling
Both MCP and the REST API return structured errors that AI tools can understand and recover from:
{
"error": "page_limit_exceeded",
"detail": "You've reached your plan limit of 100 active pages",
"hint": "Delete unused pages or upgrade your plan at /billing"
}
The error field is a machine-readable code, detail is a human-readable message, and hint tells the AI how to fix the problem. This means Claude can automatically suggest upgrading your plan or deleting unused pages when it hits a limit.
Why RunThis.page for AI publishing?
- ✓ Purpose-built for single HTML files — no build step, no Git, no deployment pipeline. Upload HTML, get a URL.
- ✓ AI-friendly errors — structured error codes let AI tools recover programmatically instead of guessing.
- ✓ Instant CDN delivery — pages are live on Cloudflare's global edge network immediately after publishing.
- ✓ Built-in analytics — track views, referrers, and trends for every page you publish.
Get started with the Claude Desktop MCP setup guide or the REST API guide. Also useful for publishing Claude artifacts and data dashboards.
Ready to share your HTML?
Upload your file and get a shareable link in seconds.
Get StartedPlans start at $15/mo. See pricing