AI Automation with MCP

Connect AI agents directly to StarLaker using the Model Context Protocol (MCP). Let Claude, GPT, or any LLM write, schedule, and publish posts — no manual steps.

What is MCP?

The Model Context Protocol lets AI models call external tools — APIs, databases, file systems — through a standardized interface. Think of it as "function calling" that works across different AI platforms.

With a StarLaker MCP server, your AI can:

• Write and publish blog posts
• Generate content on a schedule
• Cross-post to Medium and LinkedIn
• Read analytics and suggest topics
• Manage drafts and publishing queue

1. Create the MCP Server

Create a file called starlaker-mcp.js:

import { Server } from "@modelcontextprotocol/sdk/server/index.js";
import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js";

const STARLAKER_KEY = process.env.STARLAKER_KEY;
const BASE = "https://starlaker.com/api/v1";

const server = new Server(
  { name: "starlaker-mcp", version: "1.0.0" },
  { capabilities: { tools: {} } }
);

// Tool: Publish a post
server.setRequestHandler("tools/call", async (request) => {
  const { name, arguments: args } = request.params;

  if (name === "publish_post") {
    const res = await fetch(BASE + "/posts", {
      method: "POST",
      headers: {
        "Authorization": "Bearer " + STARLAKER_KEY,
        "Content-Type": "application/json",
      },
      body: JSON.stringify({
        title: args.title,
        content: args.content,
        tags: args.tags || [],
        status: args.status || "published",
      }),
    });
    const data = await res.json();
    return {
      content: [{
        type: "text",
        text: data.success
          ? "Published! " + data.url
          : "Failed: " + data.error,
      }],
    };
  }

  if (name === "list_posts") {
    const res = await fetch(BASE + "/posts?limit=10");
    const data = await res.json();
    return {
      content: [{
        type: "text",
        text: JSON.stringify(data.data, null, 2),
      }],
    };
  }

  if (name === "get_analytics") {
    // Returns top posts by views
    const res = await fetch(BASE + "/posts?limit=50");
    const data = await res.json();
    const ranked = (data.data || [])
      .sort((a, b) => (b.views || 0) - (a.views || 0))
      .slice(0, 10);
    return {
      content: [{
        type: "text",
        text: JSON.stringify(ranked, null, 2),
      }],
    };
  }

  throw new Error("Unknown tool: " + name);
});

// Register tools
server.setRequestHandler("tools/list", async () => ({
  tools: [
    {
      name: "publish_post",
      description: "Publish a new post to StarLaker. Returns the post URL on success.",
      inputSchema: {
        type: "object",
        properties: {
          title: { type: "string", description: "Post title" },
          content: { type: "string", description: "Markdown content" },
          tags: { type: "array", items: { type: "string" } },
          status: { type: "string", enum: ["draft", "published"] },
        },
        required: ["title", "content"],
      },
    },
    {
      name: "list_posts",
      description: "List recent published posts on StarLaker.",
      inputSchema: { type: "object", properties: {} },
    },
    {
      name: "get_analytics",
      description: "Get top posts by view count for content strategy.",
      inputSchema: { type: "object", properties: {} },
    },
  ],
}));

const transport = new StdioServerTransport();
await server.connect(transport);

2. Configure in Claude Desktop

Add to ~/.config/claude/claude_desktop_config.json:

{
  "mcpServers": {
    "starlaker": {
      "command": "node",
      "args": ["/path/to/starlaker-mcp.js"],
      "env": {
        "STARLAKER_KEY": "sl_live_your_key_here"
      }
    }
  }
}

Restart Claude Desktop. You'll see 🔨 tools appear in the chat.

3. Use It

In Claude Desktop, just ask:

4. VS Code Copilot Setup

Add to VS Code settings (.vscode/settings.json):

{
  "mcp": {
    "servers": {
      "starlaker": {
        "command": "node",
        "args": ["starlaker-mcp.js"],
        "env": {
          "STARLAKER_KEY": "sl_live_your_key_here"
        }
      }
    }
  }
}

5. Scheduled AI Publishing

Combine with a cron job for fully autonomous content:

#!/bin/bash
# ai-publish.sh — runs daily at 9 AM

TOPIC=$(curl -s "https://newsapi.org/v2/top-headlines?category=technology&apiKey=YOUR_KEY" | jq -r '.articles[0].title')

# Ask local LLM to write a post
CONTENT=$(ollama run llama3.2 "Write a markdown blog post about: $TOPIC. Keep it under 500 words.")

# Publish to StarLaker
curl -s -X POST https://starlaker.com/api/v1/posts \
  -H "Authorization: Bearer $STARLAKER_KEY" \
  -H "Content-Type: application/json" \
  -d "{\"title\":\"$TOPIC\",\"content\":\"$CONTENT\",\"tags\":[\"ai-generated\",\"tech\"],\"status\":\"published\"}"

echo "Published: $TOPIC"

Add to crontab: 0 9 * * * /path/to/ai-publish.sh

6. Python SDK (Alternative to MCP)

If you prefer Python over MCP, a simple wrapper:

import requests, os

class StarLaker:
    def __init__(self):
        self.key = os.environ["STARLAKER_KEY"]
        self.base = "https://starlaker.com/api/v1"

    def publish(self, title, content, tags=None, status="published"):
        r = requests.post(f"{self.base}/posts", json={
            "title": title, "content": content,
            "tags": tags or [], "status": status,
        }, headers={"Authorization": f"Bearer {self.key}"})
        return r.json()

    def list_posts(self, limit=20):
        r = requests.get(f"{self.base}/posts", params={"limit": limit})
        return r.json()["data"]

sl = StarLaker()
sl.publish("AI Generated", "## Hello from Python", tags=["ai"])

Quick Reference

See also: API Reference — VS Code Setup — Getting Started