Page Serve logo

Page Serve

Community
yonatangross
page-serve

Hand a human a rendered HTML page at a stable HTTPS URL. Registers a portless route for a file or directory (https://NAME.localhost/, never a :port), prints the URL, optionally screenshots it, and tears it down on stop. Use when an agent produced a page a human is meant to open: glyph explainers, playgrounds, decision pages, docs previews. Replaces ad hoc python http.server plus alias patterns that leave servers alive.

Overview

Publisheryonatangross
Repositoryorchestkit
Skill namepage-serve
Stars
298
Forks
35
Bundled files
4
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 4 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by yonatangross on GitHub. Read the source before you install it.

Installation

Install the Page Serve AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/yonatangross/orchestkit/tree/main/plugins/ork/skills/page-serve
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/yonatangross/orchestkit.git /tmp/orchestkit
mkdir -p .claude/skills
cp -r /tmp/orchestkit/plugins/ork/skills/page-serve .claude/skills/page-serve
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Page Serve in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Page Serve on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Page Serve is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

page-serve, give a human a URL

An agent that renders a page for a human needs a way to hand it over. The pattern that grew in sessions was python3 -m http.server 8991 plus portless alias <name> 8991 typed by hand: no record of the process, no stop, servers outliving the session, and one measured night (2026-09-06) where the same session did it twice and the server was still up the next morning. This skill is that pattern with a state file, a stop, a status, and a loud failure when the URL contract cannot be met.

The contract

  • The URL is https://<name>.localhost/<file>. No port, ever. portless's 443 service owns the port; a :1355 or :8991 in the URL is the pre-service fallback and the skill refuses rather than emit one.
  • One static server per name, on a free loopback port, rooted at the file's directory (so sibling assets resolve) or at the directory you pass.
  • State in .claude/state/page-serve/<name>.json: root, file, port, pid, url, time. status re-measures pid, route, and HTTP on every call; it never trusts the file.
  • Idempotent per name: same name + same root reuses the live server; a different root replaces it; --force on the alias overrides a stale route with the same name.

Usage

page-serve docs/playgrounds/roadmap-2026-09-07.html
page-serve docs/playgrounds/ --name playgrounds
page-serve <path> --screenshot          # agent-browser, returns the png path
page-serve status [--json]
page-serve stop <name>
page-serve stop --all

serve.sh prints url=, name=, port=, pid=, http=, optional screenshot=, and the exact stop= line to paste. Put the url= line in your reply's Open section.

Modes

InvocationScriptWhat happens
<path> [flags]scripts/serve.shprerequisites, server, alias, state, curl check, optional screenshot
status [--json]scripts/status.shone line per served page: server up/dead, route registered/missing, HTTP code
stop <name> or stop --allscripts/stop.shalias removed, server killed, state and screenshot deleted; every step reports

Failure modes, all loud

ConditionExitWhat it says
portless not installed2install line, then portless service install
443 service not responding2portless service status, the operator's launchctl kickstart line, and "do not fall back to a :port URL"
server did not bind3port and log path
alias registration failed2portless's own message, server cleaned up
route up but URL not 2004code, curl rc, and portless list; server left running for inspection
screenshot failed0reported on stderr, page still served, screenshot= line omitted

A page that "served" but does not answer 200 through the proxy is exit 4, not 0, so a caller that only reads the exit code cannot report a dead link as done.

Screenshot

With agent-browser installed the default is to take one (--no-screenshot to skip; --screenshot to insist and get a stderr line if it is missing). The png lands beside the state file. The screenshot is evidence the page rendered, not proof it rendered correctly: read the image before claiming the page looks right (the glyph skill's rule 1).

What this is not

  • Not a dev server. For a framework app with hot reload use dev, which wraps the app's own dev command in portless. This skill serves static files.
  • Not a publisher. Nothing leaves the machine; .localhost resolves locally only.
  • Not a docs-site deploy. Lab pages under docs/site/public/lab/ still ship through the docs-site build; this skill is for the hand-over before or beside that.

Retiring the old pattern (#3900)

Any skill or page that tells an agent to run python3 -m http.server or to type portless alias by hand should say page-serve <path> instead. As of this skill's first release, grep -rln 'http.server' src/skills is empty; the pattern lived in session habits and in hq-ext, so the sweep is glyph, playground, visualize-plan and expect (the four that hand a human a page) plus hq-ext's glyph.

Related skills

  • dev, the dev-loop sibling for framework apps (portless wrapping a dev command)
  • portless (the reference skill, model-invoked): service install, LAN mode, gotchas
  • glyph, visualize-plan, playground, the producers of pages this serves
  • expect, browser verification against the URL this prints

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Page Serve AI skill do?

Hand a human a rendered HTML page at a stable HTTPS URL. Registers a portless route for a file or directory (https://NAME.localhost/, never a :port), prints the URL, optionally screenshots it, and tears it down on stop. Use when an agent produced a page a human is meant to open: glyph explainers, playgrounds, decision pages, docs previews. Replaces ad hoc python http.server plus alias patterns that leave servers alive.

Why use Page Serve on TypingMind?

Because you install it once and use it with any model. Page Serve is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Page Serve in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/yonatangross/orchestkit/tree/main/plugins/ork/skills/page-serve. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Page Serve?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Page Serve?

As many as you like. As long as a model supports skills, you can use Page Serve with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Page Serve AI skill free?

Yes. It is published on GitHub by yonatangross under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇