MCP Protocol

Publish Replies via MCP Server

Give your AI assistant the ability to publish replies on Reddit and X. Works with Claude Desktop, Claude Code, Cursor, and any MCP-compatible client.

Get Your API Key Full Docs
Setup

Add to your MCP config

One JSON block in your MCP configuration file. Works with any MCP-compatible AI client.

claude_desktop_config.json 
{
  "mcpServers": {
    "dropreply": {
      "command": "npx",
      "args": ["@dropreply/mcp"],
      "env": {
        "DROPREPLY_API_KEY": "drpl_live_your_api_key"
      }
    }
  }
}
.cursor/mcp.json 
{
  "mcpServers": {
    "dropreply": {
      "command": "npx",
      "args": ["@dropreply/mcp"],
      "env": {
        "DROPREPLY_API_KEY": "drpl_live_your_api_key"
      }
    }
  }
}
Natural Language

Just ask your AI to reply

No API calls to write. Your AI agent handles the DropReply tools automatically.

Claude Desktop
You
Reply to this Reddit thread about best API tools for startups: https://reddit.com/r/SaaS/comments/abc123. Mention that we have a reply API that works great for social media automation.
Claude
I'll publish a reply to that thread using DropReply. Let me compose something helpful and relevant.

I used publish_reply to post a reply. The reply is now being published with ID rpl_7f3a9b2c. It should be live on Reddit within 30 seconds. Would you like me to check the status?
You
Yes, check it.
Claude
The reply is now published and live. Here is the direct link: https://reddit.com/r/SaaS/comments/abc123/.../def456. It was posted from an account with 12,400 karma. Would you like me to also send some upvotes?
Tools

Available MCP tools

The MCP server exposes these tools to your AI agent automatically.

publish_reply

Publish a reply to a Reddit thread or X post. Specify the platform, target URL, and reply content. Returns a reply ID for tracking.

check_reply_status

Check the current status of a published reply by ID. Returns status (queued, publishing, published, failed), platform URL, and account details.

upvote

Send upvotes to a published reply. Specify the reply ID and number of upvotes. Boosts visibility with organic-looking engagement on Reddit.

check_usage

View current usage for the billing period. Returns replies used, remaining quota, plan name, and period dates.

list_replies

List recent replies with status, platform, timestamps, and URLs. Supports filtering by platform and status for quick review.

get_reply

Get full details of a specific reply by ID. Includes content, platform URL, account info, status history, and upvote count.

Compatibility

Works with your AI tools

Any client that supports the Model Context Protocol can use DropReply.

Claude Desktop

Anthropic's desktop app with native MCP support

Claude Code

CLI coding agent with full MCP tool support

Cursor

AI code editor with MCP server configuration

Any MCP Client

Any tool supporting the open MCP protocol

Why MCP

AI-native reply publishing

Let your AI agent handle social media replies the way it handles everything else.

Natural Language

No code, no API docs. Just ask your AI to "reply to this thread" and it handles the rest using DropReply tools behind the scenes.

Secure by Design

Your API key stays in the MCP config file. The AI agent never sees the raw key - it just calls tools that are pre-authenticated.

Zero Setup Code

No SDK to install, no HTTP client to configure. Add the MCP config and your AI agent has full access to DropReply instantly.

Context-Aware

Your AI agent reads the thread, understands the context, and composes relevant replies. Better content quality than manual templating.

Full Workflow

Publish, check status, send upvotes, review usage - all from a single conversation with your AI. No context switching.

Composable

Combine DropReply MCP tools with other MCP servers. Let your AI search for threads, generate content, and publish - all in one flow.

Give your AI the power to reply

Get your API key, add the MCP config, and start publishing replies with natural language.

Get Your API Key Read the Docs