Encode self-contained HTML pages into shareable URLs.
page.anew/anew MCP Server
The page.anew/anew Model Context Protocol (MCP) server encodes self-contained HTML documents into shareable URLs and can decode them back. Its hosted MCP endpoint is at https://anew.page/mcp. It is designed to operate without accounts, databases, or deploy steps for the encoded content.
π οΈ Key Features
Provides 2 tools (toolCount: 2): anew_write and anew_read
anew_write: encodes a self-contained HTML document into an anew.page URL
anew_read: takes an anew.page URL and outputs the HTML encoded in it
Tools are read-only; anew_write mints a new URL and preserves the original
π Use Cases
Share or transport self-contained HTML via URLs
Convert between HTML and an encoded anew.page link
β‘ Developer Benefits
Works with the Model Context Protocol via https://anew.page/mcp
Supports HTML workflows using anew_write/anew_read
Enables encoding/decoding through URL inputs and outputs
β οΈ Limitations
Only two tools are provided, both described as read-only in behavior
The repository βholds the plugin, extension and skill manifestsβ and βcontains no product sourceβ
Put markup and inline CSS in html. For calculations, inputs, buttons, or other interactivity, put raw JavaScript in javascript. Never put <script> tags or inline event handlers in html, and never wrap javascript in <script> tags; the server inserts it before </body>. Encode a self-contained HTML document (or fragment, or plain text) into an anew.page URL and return that URL. Each encoding path is deterministic: the same input through the same path always produces the same URL. Different paths use different Brotli encoders, though, so the same HTML may mint different β equally valid β URLs; compare pages by their decoded HTML, never by slug. Inline all CSS, JS, and images (data: URIs, inline SVG, CSS gradients, or emoji) β the page should render alone forever, and each external reference gambles it on someone else's host staying up. One accepted exception: a webfont is far too large to inline, so a font may load from a durable CDN with a system-stack fallback; do not fight a user who asks for one. The HTML should stay β€ 160000 UTF-8 bytes and the encoded URL β€ 4000 bytes β the recommended share-safe budgets, not walls: a page past either still mints, renders and serves, up to the physical ceilings (65000 URL bytes, where the edge severs the request line; 163839 decoded bytes, the server's decode cap), it may just break in some apps β the response says so with a warning. Past a ceiling the endpoint refuses with an actionable error (code url_over_budget or content_too_large). Brotli slugs (the endpoint, or a local brotli mint) put a typical page in a few hundred URL bytes; raw base64 fits only ~2900 HTML bytes share-safe β know which budget you are drafting to, and never cut page content to fit before checking the brotli path. Returns the URL as plain text β return it to the user verbatim. To revise an existing anew page, its URL is its source: read it, edit that HTML, write it again. The new URL is a separate page and the original keeps working. Read it with anew_read first, then call this with the edited HTML.
Parameters2
html
string
required
Self-contained HTML markup and inline CSS only; no script elements, inline event handlers, or executable JavaScript. Encoded into the returned URL.
javascript
string
optional
Optional raw client-side JavaScript. Populate for any interactive behavior. Source only β no script tags. The server inserts it immediately before </body> (or at the end when there is no body).
Raw schema
{
"type": "object",
"properties": {
"html": {
"type": "string",
"minLength": 1,
"description": "Self-contained HTML markup and inline CSS only; no script elements, inline event handlers, or executable JavaScript. Encoded into the returned URL."
},
"javascript": {
"type": "string",
"minLength": 1,
"description": "Optional raw client-side JavaScript. Populate for any interactive behavior. Source only β no script tags. The server inserts it immediately before </body> (or at the end when there is no body)."
}
},
"required": [
"html"
],
"additionalProperties": false
}
anew_read
Pass an anew.page URL, get back the HTML document encoded in it, plus its canonical URL and byte size. (Two in-page tools share this name and take no arguments: the editor's reads the document open in the editor, and a rendered page's reads the page being viewed. This one takes a url and reads that page.) To revise an existing anew page, its URL is its source: read it, edit that HTML, write it again. The new URL is a separate page and the original keeps working. Read returns the author's exact bytes β no anew-injected tags, no proxied image srcs β so it re-encodes byte-for-byte. Never re-author from a rendered page, a scraped DOM, or a screenshot: each loses the author's markup, and none is faster than reading the slug. Decoding is local and exact, so any anew URL reads β the page needs no prior visit and no fetch.
Parameters2
url
string
required
An anew.page URL, or a bare slug. Subdomain, cosmetic-prefix and #hash forms all resolve.
format
string
optional
How the document rides back. "text" (default) is one plain text block. "resource" is one embedded-resource block ({type:"resource"}, mimeType text/html) β the typed MCP form for callers that want the page tagged as HTML rather than loose text. Same bytes both ways.
Raw schema
{
"type": "object",
"properties": {
"url": {
"type": "string",
"minLength": 1,
"description": "An anew.page URL, or a bare slug. Subdomain, cosmetic-prefix and #hash forms all resolve."
},
"format": {
"type": "string",
"enum": [
"text",
"resource"
],
"default": "text",
"description": "How the document rides back. \"text\" (default) is one plain text block. \"resource\" is one embedded-resource block ({type:\"resource\"}, mimeType text/html) β the typed MCP form for callers that want the page tagged as HTML rather than loose text. Same bytes both ways."
}
},
"required": [
"url"
],
"additionalProperties": false
}
anew.page makes free webpages that live in the URL. Give it one self-contained
HTML file and you get back a link that carries the whole page. No account, no hosting, nothing stored.
The response body is the page's URL. No key, no signup.
Connect your agent
The MCP server is https://anew.page/mcp: HTTP transport, no auth, two tools. anew_write takes
HTML and returns a URL; anew_read takes a URL and returns the HTML. Add the server URL to any client
that accepts one, or install this repository where your client has a plugin, extension or skill
marketplace. The one-click installs are at anew.page.
The skill is the full how-to for an agent writing pages. The
reference covers the HTTP, MCP and A2A interfaces, limits and local
minting.
Rules of a page
Inline everything. CSS, JS and images go in the file (data: URIs, inline SVG). Webfonts may load from a CDN.
4000-byte URL limit. Pages are Brotli-compressed into the link; past the limit no link is issued and the error says by how much. Post to measure, don't estimate.
Links are immutable. To edit, read the page, change the HTML, write it again. The new link is a new page; the old one keeps working.
This repo holds the plugin, extension and skill manifests and server.json (the MCP registry entry),
all pointing at https://anew.page/mcp. No product source.
Install
Tools (2)
anew_writeanew_read
Remote endpoint
Streamable HTTP
Hosted server - connect over the network, no local install.