Claude Tutorials

Tutorial: DM Management, Export & Automation with Claude

You are my X/Twitter DM management specialist. I want to use XActions to send DMs, manage conversations, export message history, and set up automated responses. Help me take control of my inbox.

Context

I'm using XActions (https://github.com/nirholas/XActions), an open-source X/Twitter toolkit with DM capabilities via MCP tools (x_send_dm, x_get_conversations, x_export_dms) and browser console scripts (src/sendDirectMessage.js, src/automation/customerService.js).

What I Need You To Do

Part 1: DM Overview

First, help me understand my DM landscape:

Get my conversations using x_get_conversations:
```
"Show me my recent DM conversations"
```
- List all active conversations
- Show unread count
- Identify important vs spam DMs
Export DM history using x_export_dms:
```
"Export my last 100 DM messages to JSON"
```
- Full message history with timestamps
- Useful for backup or analysis
- Identify patterns in who DMs you

Part 2: Sending DMs

Send direct messages through Claude:

Single DM using x_send_dm:

"Send a DM to @username saying: Hey! Loved your thread about AI. Would love to connect!"

Parameters:

username: Recipient (without @)
message: Message text

DM templates for common scenarios:

Networking:

"Hey [name]! Really enjoyed your recent thread about [topic]. Your point about [specific insight] really resonated. Would love to continue the conversation — I'm working on [your project] which is in a similar space."

Collaboration:

"Hi [name]! I've been following your work on [their project] and think there could be some great synergy with what I'm building at [your project]. Would you be open to a quick chat about potential collaboration?"

Thank you:

"Hey [name], just wanted to say thanks for [specific action — retweeting, sharing, mentioning]. Really appreciate the support! 🙏"

Cold outreach:

"Hi [name]! I'm [your name], working on [brief description]. I came across your work through [how you found them] and was impressed by [specific thing]. I'd love to get your thoughts on [specific question]. No pressure — just admire your perspective."
DM etiquette tips:
- Always personalize — never send identical messages
- Reference something specific about them
- Keep it short (2-3 sentences max for first DM)
- Don't pitch immediately
- Don't ask for follows or retweets

Part 3: Browser Console DM Script

Walk me through src/sendDirectMessage.js:

How it works:
- Navigates to X DM interface
- Opens a new conversation or existing one
- Types and sends the message
- Uses X's DM selectors:
  - [data-testid="dmComposerTextInput"] for the text input
  - [data-testid="dmComposerSendButton"] for send
Batch DM sending:
- Send personalized DMs to a list of users
- Important: Personalize each one (X flags identical mass DMs)
- Rate limit: 1-2 DMs per minute max
- Daily limit: Stay under 50 new DMs to avoid restrictions

Part 4: Customer Service Bot

For business accounts, set up src/automation/customerService.js:

Paste core.js, then configure customerService.js:

const CONFIG = {
  ACTIVE_ACCOUNT: 'your_business_account',
  
  MONITOR: {
    mentions: true,      // Watch @mentions
    dms: true,           // Watch DMs
    replies: true,       // Watch replies to your posts
    keywords: true,      // Watch brand keywords
  },
  
  BRAND_KEYWORDS: ['your_brand', 'your_product'],
  
  RESPONSE: {
    autoReply: true,
    autoReplyDM: false,  // Careful with auto DM replies
    requireApproval: true,  // Review before sending (recommended!)
    markAsRead: true,
  },
  
  BUSINESS_HOURS: {
    enabled: true,
    start: 9,   // 9 AM
    end: 17,    // 5 PM
    days: [1,2,3,4,5],  // Mon-Fri
    outsideHoursMessage: "Thanks for reaching out! We're outside business hours (Mon-Fri 9-5 ET). We'll respond next business day! 🙏",
  },
  
  CHECK_INTERVAL_SECONDS: 60,
  MAX_RESPONSES_PER_HOUR: 20,
  MAX_RESPONSES_PER_DAY: 100,
};

Response templates — the bot auto-categorizes messages:

const TEMPLATES = {
  greeting: ["Hi {customer}! Thanks for reaching out. How can I help? 😊"],
  issue: ["Hi {customer}, sorry about the trouble. Please DM us details so we can fix this."],
  thanks: ["Thank you {customer}! We really appreciate the kind words! ❤️"],
  question: ["Great question {customer}! Here's what I can tell you: [CUSTOMIZE]"],
  pricing: ["Hi {customer}! For pricing, check our website or DM us for a quote! 💰"],
  support: ["Hi {customer}! For support, DM us or email support@company.com 📧"],
};

Keyword auto-detection — Categorizes incoming messages:
- Issue keywords: problem, broken, doesn't work, refund, cancel
- Thanks keywords: thank, awesome, amazing, love, great
- Pricing keywords: price, cost, how much, discount
- Question keywords: ?, how do, what is, where, when
How it runs:
- Checks notifications every 60 seconds
- Detects new mentions/replies
- Auto-categorizes the message
- Picks appropriate template response
- If requireApproval is true, shows you the response before sending
- Respects business hours
- Rate limits itself

Part 5: DM Export & Backup

Export your entire DM history:

Via MCP:
```
"Export all my DMs to JSON"
```
Via browser script (scripts/scrapeDMs.js):
- Open X DM inbox
- Script scrolls through all conversations
- Extracts messages with timestamps, sender, and content
- Exports as JSON

Via CLI:

xactions dms --limit 500 --format json > my-dms-backup.json

Use cases for DM exports:
- Backup important conversations
- Analyze common questions (for FAQ creation)
- Track business lead conversations
- Legal/compliance records

Part 6: DM Manager (Puppeteer)

For programmatic Node.js usage, the dmManager module:

import { dmManager } from 'xactions';

// Access DM management functions:
// - Send messages
// - Read conversations
// - Export conversations
// - Manage message requests

Walk me through the available functions and how to integrate them into custom scripts.

Part 7: DM Outreach Campaign

Help me set up a tasteful DM outreach campaign:

Define the campaign:
- Who am I reaching out to? (List of target usernames)
- What's the goal? (Networking, sales, collaboration, community building)
- What's the message? (Must be personalized per person)
Research each target using x_get_profile:
- What do they tweet about?
- What's in their bio?
- Any recent achievement I can reference?

Craft personalized messages — Give me a template but help me customize each one:

For @user1 (AI researcher): "Hey [name]! Your paper on [topic] was fascinating..."
For @user2 (Startup founder): "Hi [name]! I saw your launch of [product]..."
For @user3 (Community builder): "Hey [name]! Love what you're doing with [community]..."

Send with appropriate pacing:
- 2-3 DMs per hour max
- Never more than 15-20 new DMs per day
- Wait for responses before sending to similar accounts
- Track who responded and who didn't
Follow-up strategy:
- If no response after 3-5 days, one gentle follow-up
- If still no response, move on
- Never send more than 2 messages without a reply

Part 8: DM Analytics

Track your DM performance:

Response rate: What % of your outreach DMs get responses?
Response time: How quickly do people reply?
Conversion: How many DM conversations turned into real connections?
Patterns: Which DM opener style works best?

My DM Goals

(Replace before pasting)

Purpose: Networking / Customer service / Outreach / Backup
Target users to DM: @user1, @user2 (or a description)
Am I a business account? Yes/No
Brand keywords to monitor: brand1, brand2

Start with Part 1 — show me my DM landscape and help me organize my inbox.

⚡ Explore XActions

100% free and open-source. No API keys, no fees, no signup.

Browse All Documentation