Your assistant. Your content.
Visibly Plugins.

Download and install Visibly plugins for OpenAI Codex, Claude Code and GitHub Copilot CLI. Write articles, improve NSS and save drafts to the editor.

Plugins v1.0.1 MCP v0.14.0 · 88 Tools

Choose your assistant

Download for free or install from the public Visibly marketplace on GitHub. The remote connection needs no Python installation.

OpenAI Codex

Write articles, measure NSS and improve your draft.

MCP connection included

Download ZIP Installation guide

Claude Code

Write articles, measure NSS and improve your draft.

MCP connection included

Download ZIP Installation guide

GitHub Copilot CLI

Write articles, measure NSS and improve your draft.

Connect MCP separately once

Download ZIP Installation guide

Release notes and SHA-256 checksums · Public GitHub marketplace; not listed in the providers’ curated stores.

What does NSS measure? Criteria, transparency and limits

Your assistant writes. Visibly helps measure and improve.

1. Choose a project and briefing
The agent uses your existing requirements, terms, headings and links from Visibly.
2. Write and improve NSS
The content-nss-optimize skill measures the actual NSS after each revision. Aim for 70 or 80, with a limited number of improvement rounds.
3. Continue in the editor
Get the saved draft’s editor link and measured score. Publishing to your website is a separate step.

0 Visibly credits for existing context, NSS scoring and saving drafts. Writing uses your assistant’s model; its tokens or subscription are billed by that provider. New analyses may cost credits separately. A specific NSS result is not guaranteed.

More skills and tools are available through the current MCP catalog, depending on your API key and project permissions. Explore MCP capabilities

Connect in a few steps

You need a current client with plugin support and a Visibly account with access to your project. Create your API key in Visibly settings. Set it in the same terminal before starting your client:

macOS / Linux

export VISIBLYAI_API_KEY='YOUR_VISIBLY_API_KEY'

Windows PowerShell

$env:VISIBLYAI_API_KEY = 'YOUR_VISIBLY_API_KEY'

After installing the plugin, start codex or claude from this same terminal. Copilot CLI needs the separate MCP setup below. Keep your key private and out of plugin files and Git.

On every OS, a new chat does not load a newly saved key into a running app. Fully quit VS Code, load the key in a standalone terminal and restart the editor there with code . The integrated terminal does not update the extension’s environment. The plugin does not automatically load a project .env file.

macOS

Quit VS Code with Cmd+Q; closing a window may leave the app running. Load the key in Terminal and launch with code . If the command is missing, use VS Code’s Command Palette: “Shell Command: Install code command in PATH”. A Dock or Finder launch does not automatically inherit a terminal export.

For persistence, store the key in a private file at ~/.config/visibly/env, restrict access to your account and explicitly load it in zsh. The guide includes the steps and the entry for ~/.zshrc. macOS setup

Linux

Exit all VS Code instances and run code . from the terminal containing the key. The app menu does not automatically inherit a terminal export. For future terminals, load a private key file from ~/.bashrc or ~/.zshrc; bash login shells use their login profile. Examples use bash and zsh; if you use fish, start bash for these steps. Linux setup

Windows

Save VISIBLYAI_API_KEY in your Windows user environment variables. Exit all VS Code windows, open a standalone PowerShell and load the saved value with the command below. Then run code . from your project directory. A new chat or Reload Window is not sufficient.

$env:VISIBLYAI_API_KEY = [Environment]::GetEnvironmentVariable('VISIBLYAI_API_KEY', 'User')

Windows setup and key check

For WSL, SSH or containers, the key must be available in the environment where the MCP client runs.

OpenAI Codex

codex plugin marketplace add AntonioBlago/visiblyai-mcp-server
codex plugin add visibly-codex@visibly

Claude Code

claude plugin marketplace add AntonioBlago/visiblyai-mcp-server
claude plugin install visibly-claude@visibly

GitHub Copilot CLI

copilot plugin marketplace add AntonioBlago/visiblyai-mcp-server
copilot plugin install visibly-copilot@visibly

This plugin includes the skill. Connect MCP separately once: run /mcp add in Copilot, choose HTTP, and enter the URL below with an Authorization header containing Bearer followed by your API key.

https://mcp.visibly-ai.com/mcp

These commands download the plugins automatically. For a manual installation, extract the ZIP and follow the client-specific instructions. Full guide on GitHub

Verify the connection

After restarting, open a new chat and ask:

Call list_projects through the Visibly MCP and show my available projects. Report connection or permission errors. Do not start an analysis or modify content.

This read costs 0 Visibly credits. A successful empty list means the connection works but no projects are visible to this key. A visible skill alone does not verify the connection. For missing tools, check the launch environment; for 401 check the key, and for 403 check permissions and project access. After rotating a key, also update Copilot’s saved header.

Your first request to the assistant

Use content-nss-optimize for my Visibly project and ready content briefing. Write with your own model, optimize toward NSS 80, save the draft and give me the editor link. Do not use paid Visibly text generation.

From your AI assistant to your CMS

The plugins connect your assistant to Visibly: read the briefing, write, measure NSS and save a draft in the editor. The Visibly AI CMS Connector connects your Python application to the article API: receive a webhook, fetch the content and confirm the published URL.

anyCMS provides concrete use cases for WordPress, Astro, Next.js and Flask. A saved draft or an accepted webhook does not yet mean the article is published on your website.

Visibly AI CMS Connector anyCMS: use cases CMS setup

Claude Desktop, ChatGPT or VS Code?

These downloads target Claude Code, OpenAI Codex and Copilot CLI. Claude Desktop and other compatible clients can connect MCP separately. Copilot in VS Code needs its own MCP and skill setup. These plugins do not include a direct ChatGPT integration with OAuth login. MCP setup

Also available: SEO Starter for Claude Code

A separate plugin for SEO consulting: status quo, potential, proposals and CI PDFs.

Quick Start
Included Commands
Skills
What Gets Connected
Requirements
Source Code and License

Quick Start

The plugin lives in a public marketplace on GitHub. Add the marketplace once, then install the plugin. Both steps run directly in Claude Code:

# 1. Add the marketplace
/plugin marketplace add AntonioBlago/claude-code-seo-starter

# 2. Install the plugin
/plugin install seo-starter@antonioblago

Add the marketplace

Registers the antonioblago marketplace in your Claude Code installation. One-time per machine.

Install the plugin

Loads the seo-starter plugin with all commands, skills and the MCP configuration.

Add your API key

Set your Visibly AI key as the environment variable VISIBLYAI_API_KEY to activate the MCP tools.

You can get your free API key in your account under MCP Server > Setup. The workflow commands also work without a key, but live SEO data only comes through once you add one.

Included Commands

Four slash commands cover the complete SEO consulting workflow, from status quo to the finished proposal PDF. Each command automatically calls the matching Visibly AI tools and skills.

Command	Purpose
`/visibly-seo-status-quo`	Visibility analysis of a domain in the target market: GSC data, keyword classification, quick wins.
`/visibly-seo-potential`	Potential analysis using the intent-based CTR model from the Keyword Study 2026: traffic and lead potential per cluster.
`/visibly-seo-offer`	Tailored SEO proposal based on the analysis data: phases, services, investment, ROI.
`/visibly-seo-pdf-build`	Generates a CI-compliant PDF from the proposal using Montserrat and Aptos fonts.

Skills

Skills are knowledge packages that Claude Code loads automatically when a command needs them. You don't have to call them manually, they provide the domain context in the background.

Skill	Content
`seo-status-quo`	Methodology for the visibility analysis: GSC queries, keyword cross-referencing, status classes.
`seo-potential-analysis`	CTR model, target positions, delta calculation, lead and ROI calculation.
`seo-offer`	Structure of a proposal: phase model, deliverables, management summary.
`ci-pdf-build`	CI guidelines for PDFs: colors, fonts, layout, table logic.

What Gets Connected

The plugin includes a ready-made MCP configuration that connects the Visibly AI server as a remote endpoint. This gives you all SEO tools directly in Claude Code: Search Console, keyword research, backlinks, competitor benchmarking, on-page audits and crawls.

// .mcp.json (included in the plugin)
{
  "mcpServers": {
    "visiblyai": {
      "type": "http",
      "url": "https://mcp.visibly-ai.com/mcp",
      "headers": {
        "Authorization": "Bearer ${VISIBLYAI_API_KEY}"
      }
    }
  }
}

The placeholder ${VISIBLYAI_API_KEY} is filled from your environment variable. Your key never ends up in the plugin code. You'll find details on all tools and credit costs in the MCP Documentation.

Requirements

Claude Code (CLI, Desktop, web or IDE extension) in a current version.
A GitHub account isn't required, the marketplace is public.
For live data: a free Visibly AI API key as the environment variable VISIBLYAI_API_KEY.
For PDF export: a local Python environment with fpdf2 (described by the skill).

Source Code and License

The plugin is open source under the MIT license. You can view the full code, fork it and adapt it for your agency.

View repository, commands and skills on GitHub

Open GitHub repository

Choose your assistant

OpenAI Codex

Claude Code

GitHub Copilot CLI

Your assistant writes. Visibly helps measure and improve.

1. Choose a project and briefing

2. Write and improve NSS

3. Continue in the editor

Connect in a few steps

OpenAI Codex

Claude Code

GitHub Copilot CLI

Verify the connection

Your first request to the assistant

From your AI assistant to your CMS

Claude Desktop, ChatGPT or VS Code?

Table of Contents

Quick Start

Add the marketplace

Install the plugin

Add your API key

Included Commands

Skills

What Gets Connected

Requirements

Source Code and License