Bulk Publishing AI MCP Server — Overview

What Is the MCP Server?

The Bulk Publishing AI MCP (Model Context Protocol) Server turns the entire content generation platform into a set of callable tools that any MCP-compatible AI client can access. Instead of using the web dashboard, site owners can generate articles, manage projects, publish to WordPress, export content, and trigger SEO indexing directly from their AI coding environment.

Compatible Clients:

  • Antigravity
  • Claude Desktop
  • Cursor
  • Windsurf
  • Cline
  • Any MCP-compatible editor or agent framework

Architecture

┌─────────────────────────────┐
│   MCP Client (Claude, etc.) │
│   User sends natural        │
│   language commands          │
└──────────┬──────────────────┘
           │ stdio / SSE
           ▼
┌─────────────────────────────┐
│   BPAI MCP Server           │
│   ┌───────────────────────┐ │
│   │  Tool Router          │ │
│   │  ┌─────────────────┐  │ │
│   │  │ generate         │  │ │
│   │  │ research         │  │ │
│   │  │ export / import  │  │ │
│   │  │ wordpress        │  │ │
│   │  │ indexing          │  │ │
│   │  │ projects          │  │ │
│   │  └─────────────────┘  │ │
│   └───────────────────────┘ │
│              │               │
│   ┌──────────▼────────────┐ │
│   │  AI Engine             │ │
│   │  OpenAI | Claude |     │ │
│   │  Gemini | Grok |       │ │
│   │  DeepSeek | Kimi |     │ │
│   │  Perplexity            │ │
│   └──────────┬────────────┘ │
└──────────────┼──────────────┘
               │
    ┌──────────▼──────────┐
    │  InsForge Backend    │
    │  PostgreSQL + Auth   │
    │  + Edge Functions    │
    └──────────┬──────────┘
               │
    ┌──────────▼──────────┐
    │  External Services   │
    │  WordPress REST API  │
    │  IndexNow / Bing     │
    │  Google Ping          │
    │  Broadcaster Network │
    └─────────────────────┘

Complete Tool Catalog

Content Generation

Tool Description
generate_article Generate a single SEO-optimized article from a keyword or topic
generate_batch Generate multiple articles from a list of keywords
generate_seo_metadata Generate optimized title, slug, and meta description for a keyword
humanize_content Run anti-AI detection pass on existing content

Article Library

Tool Description
list_articles Browse and search your generated article library
get_article Retrieve full content of a stored article by ID

Research

Tool Description
research_topic Run web research on a topic using Perplexity or other providers

Import & Export

Tool Description
import_csv Import keywords, titles, and metadata from a local CSV file
export_csv Export articles as CSV
export_markdown Export articles as individual .md files with YAML frontmatter
export_json Export articles as structured JSON
export_bulk Combined export with format, scope, and filter options

WordPress Publishing

Tool Description
publish_to_wordpress Publish a single article to a connected WordPress site
publish_batch Publish multiple articles in sequence with rate limiting
test_wp_connection Verify WordPress credentials and connectivity
list_wp_categories List available categories from a WordPress site

SEO & Indexing

Tool Description
broadcast_urls Submit URLs to Google, Bing, and Yandex discovery endpoints
indexnow_submit Submit URLs via the IndexNow protocol for rapid indexing
fetch_sitemap Fetch and parse a site's XML sitemap

Project Management

Tool Description
list_projects List all projects for the authenticated user
switch_project Set the active project for subsequent commands
get_project_settings View settings for a project (WordPress, linking, prompts)
get_settings View current AI generation settings
update_linking_settings Update internal linking configuration for a project

Knowledge Base

Tool Description
list_knowledge_base List knowledge base items for a project
add_knowledge_base Add brand voice, facts, or context snippets
update_knowledge_base Update an existing KB item
delete_knowledge_base Remove a KB item

Models

Tool Description
list_models List available AI models for a provider

Image Generation

Tool Description
generate_image Generate a single image from a text prompt using Gemini Nano Banana models
generate_article_images Generate hero + section images for an article keyword
list_image_models List available Gemini image models and capabilities

MCP vs. Web App Comparison

Capability Web App MCP Server
Single article generation ✅ ✅
Batch generation ✅ (browser-limited) ✅ (server-side)
CSV input Upload (10MB limit) Local file read (unlimited)
CSV/Markdown/JSON export Browser download Direct file write
WordPress publishing ✅ ✅
Deep Research ✅ ✅
Internal Linking ✅ ✅
Cross-Linking ✅ ✅
Broadcaster / IndexNow ✅ ✅
Knowledge Base ✅ ✅
Article Persistence ✅ ✅
Scriptable / automatable No Yes
Works without browser No Yes
File system access No Yes
Image generation Yes Yes

BYOK (Bring Your Own Key) Model

The MCP server uses the same BYOK architecture as the web app. Users provide their own API keys for AI providers:

  • OpenAI — GPT-5.2, GPT-5-mini, o4-mini, GPT-4.1 series
  • Anthropic — Claude Opus 4.5, Claude Sonnet 4.5, Claude 3.7 Sonnet
  • Google — Gemini 3 Pro, Gemini 2.5 Flash
  • xAI — Grok 4 Flagship, Grok 4.1 Fast
  • DeepSeek — DeepSeek Reasoner V3.2
  • Kimi — Kimi K2.5 Thinking
  • Perplexity — Sonar Pro, Sonar Reasoning Pro

Multi-key registration allows parallel generation across multiple API keys per provider, distributing rate limits and maximizing throughput.

Security Model

  • Row Level Security (RLS): All data is isolated per user via auth.uid() = user_id
  • API Key Encryption: User API keys are stored encrypted in InsForge PostgreSQL
  • No Key Storage on MCP: The MCP server never persists API keys locally. Keys are fetched per-session from the database
  • TLS Transit: All communication between the MCP server and InsForge uses encrypted connections
  • Project Isolation: Each project's settings, knowledge base, and content are isolated by project_id

Next Steps