Guides
Publishing what your AI agent makes
Short, practical guides for getting what your coding agent produces — reports, dashboards, HTML pages, PDFs — out of your terminal and into a link you can share.
Coding agents like Claude Code, Codex, Cursor and Windsurf are great at making things. The gap is everything after "done": the output lands as a file in a folder, and the moment someone who doesn't live in your terminal needs to see it, the options are all a little wrong — email an attachment, screenshot away the interactivity, or spin up hosting for a page you'll share once.
These guides cover the practical middle ground: how to turn agent output into a real URL with the right people able to open it — and an honest look at the alternatives (GitHub Pages, S3, Netlify, a gist) so you can tell when freo.cloud is the fast path and when it isn't. Each one is written for the moment you actually have the file in hand.
They show the coding-agent path — the publishing skill (Claude Code, Codex, Cursor, Windsurf and more). On a chat assistant like Claude.ai or ChatGPT? You publish the same way over the MCP server — one-click OAuth, no token to paste.
In a hurry? Connect first, read later
Claude Code
Install the plugin, then /mcp → freo → Authenticate:
/plugin marketplace add nanosolutions/freo-skills /plugin install freo-publish@freo
Claude.ai · ChatGPT · Claude Desktop
Claude: Settings → Connectors → add this URL, then sign in. ChatGPT: Settings → Apps → Advanced settings → turn on Developer mode, then Create with this URL (paid plans).
https://freo.cloud/mcp
Codex, Cursor, opencode, Gemini CLI… · any other coding agent
Paste this into your agent. It installs the publishing skill, then asks for the one-time setup link from your Connect page:
Install the freo.cloud publishing skill from https://freo.cloud/skill and use it to publish my documents.
Start here
How to host an HTML file or dashboard from your AI agent
Got a self-contained HTML page or an interactive dashboard from your agent? Here is how to get it a real URL — interactive, private, no build pipeline.
Read the guide →Access control for AI-generated files
Public, unguessable link, email-verified, or a named allowlist — publish with the right access level, and change or revoke it any time.
Read the guide →How to tell if someone opened the report you shared
Every document tracks its own views — so you can see whether the client actually opened the link before you chase them.
Read the guide →How to share Claude Code output as a link
Turn a Claude Code report, dashboard, HTML page or PDF into a shareable link — without spinning up hosting.
Read the guide →Where to put the files your AI agent makes
GitHub Pages, S3, a gist, or freo.cloud? A fair comparison — and how to get one shareable link.
Read the guide →How to make a share link that expires on its own
Give a shared document a hard expiry, or let it auto-retire after nobody opens it — and bring it back in one click.
Read the guide →How to catch leaked API keys before you share AI output
AI agents paste real credentials into examples and logs. Here is how every publish is scanned for secrets and blocked before it goes live.
Read the guide →How to send a client a link to an AI-generated deliverable
A clean, access-controlled link instead of an email attachment — locked to the client, professional, and you can see when they opened it.
Read the guide →Ready to try it?
Sign in with your email and connect your agent in one prompt — free during early access.
Get your space →