Documentation Index
Fetch the complete documentation index at: https://docs.simplehost.dev/llms.txt
Use this file to discover all available pages before exploring further.
Natural Language
Just tell your agent what to publish. The skill triggers automatically on phrases like:
- “Publish this project”
- “Deploy my dist folder”
- “Give me a live URL for this”
- “Host this HTML file”
- “Make this site live”
- “Share this as a website”
- “Update my site”
Slash Command
You can also invoke the skill directly:
Or without arguments (the skill will auto-detect what to publish):
Examples
Publish a project
You: "Publish my project to the web"
Agent: I'll publish the ./dist directory to SimpleHost.
📦 Scanning files... Found 15 files
⬆️ Uploading 15 files...
✅ Published!
🔗 https://bright-canvas-a7k2.simplehost.dev/
Update an existing site
You: "Update my site with the latest changes"
Agent: I'll update bright-canvas-a7k2 with the new build.
📦 Scanning files... Found 15 files
Skipped 12 files (unchanged)
⬆️ Uploading 3 files...
✅ Updated!
🔗 https://bright-canvas-a7k2.simplehost.dev/
First-time publish (no account)
You: "Publish this site"
Agent: 📦 Scanning files... Found 15 files
⬆️ Uploading 15 files...
✅ Published!
🔗 https://bright-canvas-a7k2.simplehost.dev/
⚠️ This link expires in 24 hours. Want me to make it
permanent? It's free — I just need your email.
You: "yes, use me@example.com"
Agent: Check your email for a code (XXXX-XXXX).
You: "ABCD-1234"
Agent: ✅ Done! Your site is now permanent.
From now on, all publishes are permanent automatically.
Publish blocked because a private key was found
You: "Publish this site"
Agent: I found a private key in your website files. If I publish it like this,
other people could see it and misuse it.
I can secure it for you before publishing, and I will not change your
original project files. Would you like me to do that?
You: "yes"
Agent: Perfect. I’ll first check whether secure browser-side secret handling is
available on this SimpleHost deployment.
If it is, I’ll secure the published version and keep your original files
untouched.
If secure browser-side secret handling is unavailable
Agent: I found a private key in your website files, and SimpleHost will not
publish it publicly in that form.
This deployment does not currently allow browser-facing secret proxying,
so I can’t safely publish this project with the secret in place.
Your original local files are still untouched.
If the user says no
Agent: I can’t safely publish this version because it contains private keys
that would be publicly visible. If you want, I can secure it first and
then publish it for you.
Ask a product question
You: "Where do I manage secrets in the dashboard?"
Agent: Open the Sites page, find your site, click the three dots on the right,
then choose Manage Secrets.
That opens the Secrets & Proxy panel for that site, where you can store
secret values and approve which external hosts the site can connect to.
You: "Why was my publish blocked?"
Agent: SimpleHost found something in your website files that looks like a
private key. Because website files are public, it blocked the publish
to keep that key from being exposed.
If secure browser-side secret handling is available, I can help secure
the published version without changing your original local files.
Auto-Detection
When you don’t specify a directory, the skill looks for common build outputs:
dist/ — Vite, Rollup, Parcel
build/ — Create React App, Webpack
out/ — Next.js static export
public/ — Static files
.next/out/ — Next.js
If none are found, the agent asks you what to publish.
Tips
Set SIMPLEHOST_API_KEY in your shell profile so the skill can publish permanently without asking for authentication each time.
- The skill remembers the slug from the last publish in the conversation, so “update my site” works without specifying the slug again
- You can publish single files too — the skill creates a temporary directory automatically
- The skill works with any file type: HTML, images, PDFs, videos, ZIP files, etc.
- If a publish is blocked because a secret was detected, the agent should explain it in plain language and ask permission before securing anything