Bloghunch MCP Server
Official# Bloghunch MCP Server
Connect your AI assistant directly to your Bloghunch publication to automate content creation, analytics, and distribution.
Supports **two transport modes**:
- **STDIO** — Claude Desktop & Cursor (local, no hosting needed)
- **HTTP** — ChatGPT (remote, requires public HTTPS URL)
---
## Tools Available
| Tool | Description |
|---|---|
| `bh_get_stats` | Analytics snapshot (pageviews, subscribers, posts) |
| `bh_list_posts` | List posts with status/keyword filter |
| `bh_create_post` | Create a post — provide `topic` for full AI generation |
| `bh_generate_ideas` | Brainstorm 5 post angles for any topic |
| `bh_list_subscribers` | List newsletter subscribers |
| `bh_social_echo` | Generate Twitter, LinkedIn, newsletter assets for a post |
| `bh_discover_topics` | Trigger GSC-powered SEO topic discovery |
| `bh_get_topic_discoveries` | List discovered topic opportunities |
---
## Installation — Claude Desktop & Cursor (STDIO)
### 1. Configure `.env`
```bash
cp .env.example .env
```
Fill in:
```
BLOGHUNCH_API_KEY=bh_live_xxxxxxxxxxxx
BLOGHUNCH_SUBDOMAIN=your-subdomain
BLOGHUNCH_API_URL=https://api.bloghunch.com/api/v1
```
Get your API key from [Settings → Developers](https://app.bloghunch.com/app/settings/developers).
### 2. Build & Auto-Install (macOS)
```bash
npm install
npm run build
npm run install-claude
```
Then restart Claude Desktop.
### 3. Manual Configuration (Windows / Cursor)
Add to your Claude Desktop config (`~/Library/Application Support/Claude/claude_desktop_config.json`) or Cursor MCP config:
```json
{
"mcpServers": {
"bloghunch": {
"command": "node",
"args": ["/absolute/path/to/mcp-server/dist/index.js"],
"env": {
"BLOGHUNCH_API_KEY": "bh_live_xxxxxxxxxxxx",
"BLOGHUNCH_SUBDOMAIN": "your-subdomain",
"BLOGHUNCH_API_URL": "https://api.bloghunch.com/api/v1"
}
}
}
}
```
---
## Installation — ChatGPT (HTTP mode)
ChatGPT requires a **publicly hosted HTTPS server**. Run this on your VPS behind Nginx/Caddy.
### 1. Build
```bash
npm install
npm run build
```
### 2. Set environment variables on your server
```bash
export TRANSPORT=http
export PORT=3001
export MCP_SERVER_URL=https://mcp.yourdomain.com # public HTTPS URL
export BLOGHUNCH_API_URL=https://api.bloghunch.com/api/v1
```
### 3. Start the server
```bash
npm start
# or: TRANSPORT=http node dist/index.js
```
Or with Docker:
```bash
docker build -t bloghunch-mcp .
docker run -p 3001:3001 \
-e MCP_SERVER_URL=https://mcp.yourdomain.com \
bloghunch-mcp
```
### 4. Expose via Nginx/Caddy
Point your domain (e.g., `mcp.yourdomain.com`) to the server on port 3001 with HTTPS.
**Nginx example:**
```nginx
server {
server_name mcp.yourdomain.com;
location / {
proxy_pass http://127.0.0.1:3001;
proxy_http_version 1.1;
proxy_set_header Upgrade $http_upgrade;
proxy_set_header Connection "upgrade";
proxy_set_header Host $host;
proxy_read_timeout 300s;
}
}
```
### 5. Connect in ChatGPT
1. Open ChatGPT → **Settings → Apps & Connectors → Advanced settings** → enable **Developer Mode**
2. Click **Create connector**
3. Enter name: `Bloghunch` and URL: `https://mcp.yourdomain.com`
4. Select **OAuth** as auth method
5. Click **Connect** → log in with your Bloghunch API key
6. Done ✅ — ChatGPT can now use all Bloghunch tools
---
## Local Development
```bash
# STDIO mode
npm run dev
# HTTP mode (test locally)
npm run dev:http
```
---
© 2026 Bloghunch — AI-native blogging platform.
TDQS
Scored across 8 tools
Each tool targets a distinct action and resource: stats, posts, ideas, subscribers, social echo, and topic discovery. The paired trigger/list tools for topic discoveries are clearly differentiated by their verb prefixes.
All tools share the 'bh_' prefix and use snake_case, mostly with verb_noun structure. 'bh_social_echo' deviates from the verb-first pattern, and 'get_' vs 'list_' is slightly inconsistent, but the overall convention is predictable.
Eight tools is well-scoped for a blog publication server covering analytics, content creation, subscriber management, social sharing, and SEO discovery. No bloat or redundancy.
The core workflows of creating posts, listing subscribers, gathering stats, and discovering topics are covered. Missing post update/delete and a direct 'get post' operation are minor gaps, as external dashboards likely handle those.