MCP for agents — aft.page

What it is

aft.page hosts small software — one-pagers, prototypes, agent-built HTML, static SPAs — at a shareable HTTPS URL in seconds. The name comes from the aft of a ship: the back of the vessel — take what you built and put it out into the world.

Prefer the MCP tools below. If MCP is unavailable, call the API directly.

Quick start

Thin remote MCP — one deploy tool (html or files), plus aft_deploys, aft_rollback, aft_health. No account. Point your host at:

https://mcp.aft.page/mcp

MCP host config (preferred)

Add a server named aft-page with the remote URL (Cursor, Claude Desktop, and other Streamable HTTP clients):

{
  "mcpServers": {
    "aft-page": {
      "url": "https://mcp.aft.page/mcp"
    }
  }
}

Reload the host, then ask:

Deploy this HTML to aft.page

The agent should call deploy and return a live URL.

Tools

aft.page does not build your app. Detect first: plain HTML → upload those files; Vite/React → npm run build then dist/ only; Next static export → out/. Never upload src/ or node_modules.

`deploy`

One tool. Pass html or files — not both. Goes live immediately. One-file pages use html; multi-file / SPA build output uses files (always include index.html).

Input	Type	Required	Notes
`html`	string	one of	Full document; prefer `<!DOCTYPE html>…</html>`
`files`	array	one of	Built static files. Paths relative; UTF-8 or base64.
`preferred_slug`	string	no	Hint like `about-me`. First hit never overwrites — collision gets a suffix. Required with `edit_token`.
`edit_token`	string	no	From first deploy / `.aft/state.json`. When set, MCP PATCHes this slug — same URL, new rollback row.

Returns text + structured content:

Live: https://{slug}.aft.page
Claim: https://aft.page/claim?token=…
slug: …
deploy: dep_…
files: 1
Claim within 60 minutes or drop again.

Limits: 500 files · 25 MB each · 100 MB total. Paths: no .., no leading /, no \.

`aft_deploys`

List rollback-able deploys for a slug. Needs edit_token from .aft/state.json. Same D1 history as the project UI after claim.

`aft_rollback`

Point the live slug at a prior deploy_id. Same URL. Unclaimed: edit_token. After claim, the anonymous token is retired: the owner mints one with POST /v1/sites/{slug}/deploy-token while signed in and hands it to the agent (DELETE revokes it).

`aft_health`

Ping the deploy API. No inputs. Returns ok and the API base.

Prompt: `deploy_to_aft`

Optional prompt that steers the agent:

Detect first: plain HTML vs JS app (Vite/React/Next).
Call deploy with html or files.
Vite/React: npm run build, then dist/ only. Next static export: out/.
Pass preferred_slug when sensible. Persist .aft/state.json; later deploys send edit_token.
Return the live HTTPS URL as the deliverable.
Do not invent account or repository setup steps.

Arg: html_or_files (optional context string).

API (no MCP)

Base: https://api.aft.page. CORS: * on deploy + site GET.

`GET /health`

{ "ok": true }

`POST /v1/deploy`

A) Raw HTML

POST /v1/deploy?slug=hello
Content-Type: text/html; charset=utf-8

<!DOCTYPE html><html><body><h1>Hi</h1></body></html>

B) JSON files (agents)

POST /v1/deploy
Content-Type: application/json

{
  "files": [
    { "path": "index.html", "content": "<h1>App</h1>" },
    { "path": "style.css", "content": "body{margin:0}" }
  ]
}

C) Multipart — fields file0 + file0_path, …

Success (200)

{
  "ok": true,
  "slug": "hello",
  "deployId": "dep_…",
  "url": "https://hello.aft.page",
  "files": 1,
  "bytes": 42,
  "editToken": "aft_edit_…",
  "claimUrl": "https://aft.page/claim?token=…",
  "owned": false,
  "notice": "Claim within 60 minutes or drop again."
}

Errors include no_files, too_many_files, file_too_large, payload_too_large, bad_path, reserved_slug, slug_exhausted, internal. Invalid ?slug= patterns are ignored (random unique slug); only reserved names return reserved_slug.

Serving

URL	Behavior
`https://{slug}.aft.page/`	Primary — `index.html`
`https://{slug}.aft.page/{path}`	Static asset
`https://api.aft.page/s/{slug}/`	Path fallback

If the requested object is missing, the Worker soft-falls back to that deploy’s index.html (SPA-style). Content-Type follows the file extension when the upload type is missing or application/octet-stream (so CSS is text/css).

Slugs

Prefer valid ?slug= / preferred_slug when provided.
Else allocate a random unique slug.
Clients (paste UI, extension, agents) may derive a hint from <title> / <h1> and pass it as ?slug= — the Worker does not parse HTML for a name.
First POST never overwrites — append a short word suffix on collision. PATCH with edit_token updates that slug in place.
Reserved: www, api, app, admin, docs, status, …

Claim / manage

https://aft.page/claim?token=…

The live URL is the product. After deploy, open claimUrl to add email or Google. Unclaimed sites also show Claim on the live slug. After claim, Manage opens the project page (invite, rollback, secrets, custom domains — first one free). Keep editToken to prove you deployed.

Limits (today)

Limit	Value
Files per deploy	500
Per-file size	25 MB
Total payload	100 MB
Auth	None (anonymous)
Unclaimed retention	Deleted after 30d idle. Visit, PATCH, or claim to keep.
Overwrite	Never
Custom domains / private invite	After claim · first one free (Domains tab)

Examples

Minimal HTML via curl

curl -sS -X POST "https://api.aft.page/v1/deploy?slug=curl-smoke" \
  -H "content-type: text/html; charset=utf-8" \
  --data-binary @index.html

Agent rule of thumb

Prefer MCP deploy (html or files).
Return the live URL.
Optionally link the preview shell.
Do not invent auth steps that do not exist yet.

Troubleshoot

Symptom	Fix
MCP tools missing	Reload host; check absolute path; Node 20+
`reserved_slug`	Change title / slug
Size errors	Stay under 25 MB / file, 100 MB total
Unstyled CSS	Ensure path ends in `.css`; re-deploy
Site 404	Wait for DNS; try `/s/{slug}/` on API host

MCP for aft.page