Revdoku documentation

Create websites from your AI for FREE

Ask ChatGPT, Claude or other AI to publish to Revdoku.

Get a live *.localhost3000.love website in seconds.

Free account available.

Revdoku publishes static websites and SPAs from AI-generated files and folders. Sign in to save a private draft, review a temporary 15-minute preview, and publish the main website only when you are ready.

Quick start

If npm is available, install the Revdoku skill:

npx skills add revdoku/revdoku --skill revdoku -g

Otherwise install the local client and skill:

curl -fsSL https://revdoku.com/install.sh | bash

Publish the current folder:

revdoku p

Without credentials, the client opens browser sign-in. Create a Free account there if needed. Use revdoku preview for a temporary 15-minute review URL, then run revdoku p when you want to publish. Re-running the command from the same bound folder updates the same URL.

Free account

New accounts start on the permanent Free plan. Read current prices, limits, and entitlements from Revdoku pricing or its versioned JSON contract instead of relying on copied limits. Permanent Public websites allow search indexing by default. Password, Require Email, and temporary preview websites are always noindex.

Publish a folder

Publish a specific static output directory:

revdoku p ./dist

A framework project must be built first. Publish the generated static directory, commonly dist, build, or out, rather than a development server or source-only folder.

Revdoku serves HTML, CSS, JavaScript, images, fonts, PDFs, and other static assets. It does not run a server-side application process, database, background worker, or private server environment.

Keep the same URL

A local project binding remembers which Revdoku website belongs to a folder. To update safely:

  1. Edit the original project.
  2. Run its production build if required.
  3. Publish the same bound output folder.
  4. Reload the original website URL.
  5. Confirm the revision appears.

Do not create a new website for every revision. Preserve the bound folder and reuse the existing website or bucket.

Hosted MCP

Compatible remote clients connect to:

https://app.revdoku.com/mcp

Use Streamable HTTP and browser-based OAuth. See the MCP website-publishing guide for Claude, Codex, ChatGPT, OpenClaw, and generic client instructions.

Hosted MCP works best with text-based site files and existing Revdoku websites. A remote server cannot read arbitrary files from your computer. Use the local client for folders containing local images, fonts, PDFs, archives, or other binary assets.

Access settings

Revdoku website settings can expose Public, Password, or Require Email access. Confirm the intended access option in the account before promising it to a recipient.

For Password access, send the password separately from the website link. For Require Email, remember that verifying an inbox does not prove a person’s legal identity, employer, job title, or intent.

Always test a protected website in a private browser window before sharing it.

Common commands

Command Purpose
revdoku p [folder] Publish or update a folder
revdoku ls List websites and buckets
revdoku o Open the bound website
revdoku st Show connection and account status
revdoku down Remove the bound website from public access
revdoku login Connect a different existing account

Run revdoku --help for the complete command reference.

API and security

Use the Revdoku website-publishing API guide for custom integrations. Keep credentials out of prompts, logs, repositories, and published files.

Before publishing any AI-generated site:

  • Remove secrets and private files.
  • Check external scripts and form destinations.
  • Verify every link and asset path.
  • Test mobile and desktop layouts.
  • Confirm the intended access setting.
  • Save the website identifier and URL for later updates.
Markdown version
↧
Loading PDF…