Documentation Index

Fetch the complete documentation index at: https://docs.augustin.ai/llms.txt

Use this file to discover all available pages before exploring further.

​

Installed OpenClaw… Now what? 20 use cases + All prompts

​

Overview

This comprehensive guide covers real-world OpenClaw usage based on 50+ days of daily experience, from initial ClawdBot through MoltBot to current OpenClaw. The video provides 20 practical use cases with actual prompts and addresses common challenges after the initial setup.

​

What is OpenClaw?

OpenClaw is an always-on AI agent that:

• Runs 24/7 on your server, VPS, Mac Mini, or Raspberry Pi
• Connects to existing messaging apps (Telegram, WhatsApp, Discord, iMessage)
• Functions as a persistent AI assistant accessible from anywhere

​

Evolution of Usage (50-Day Journey)

​

Week 1: Novelty Phase

• Random questions and testing
• Using like ChatGPT
• Key Decision: Markdown-first approach using Obsidian
  • Plain text files for all data
  • Human and program readable
  • No vendor lock-in
  • Easy migration to future tools

​

Week 3: Automation Building

• Daily briefings
• Background checks
• Increasing utility

​

Week 5: Context Pollution Problem

• Everything mixed in one conversation
• Research bleeding into analytics
• Bookmarks polluting daily tasks
• Solution: Separate contexts with dedicated channels

​

Week 7: System Architecture

• Different models for different tasks
• Opus for deep thinking, cheap models for routine work
• Cost optimization through model matching

​

Week 8+: From Chatbot to System

• Complete workflow integration
• Proactive assistance
• Background automation

​

Three Core Principles

1. Everything in Markdown - Future-proof, portable data
2. Separate Contexts - Dedicated channels for different workflows
3. Match Model to Task - Cost-effective model selection

​

20 Use Cases by Category

​

Automated Daily Operations

​

1. Morning News Briefing

• Function: Scans Twitter feeds at 7 AM daily
• Process: Selects top 10 tweets, writes to Obsidian, appends video ideas
• Value: Personalized news digest without scrolling
• Connection: Links current events to ongoing projects

​

2. Historical Mystery Display

• Function: Daily Wikipedia “On This Day” events
• Process: Generates woodcut-style images showing moments before historical events
• Display: TRMNL e-ink display in mystery mode (date/location only)
• Examples:
  • Iceberg approaching Titanic
  • Apple about to fall on Newton
  • Beatles rooftop concert
  • Space Shuttle Columbia disaster

​

3. System Maintenance

• 4 AM: Auto-updates OpenClaw skills from ClawHub
• 4:30 AM: Backs up all configuration files
• Includes: Workflow directory, cron schedules, SOUL files, MEMORY files, skills
• Recovery: 30-minute restoration vs. complete rebuild

​

4. Background Health Checks

• Frequency: Every 30 minutes
• Monitors: Email, calendar, services
• Catches: Payment failures, domain renewals, missed meetings, relevant content
• Email Mode: Draft-only (can read, flag, draft, but not send)

​

Research and Content Creation

​

5. Parallel Research Agents

• Process: Spawns 5 simultaneous sub-agents
  • Twitter search
  • Reddit crawling
  • Hacker News analysis
  • YouTube competition review
  • Forum scraping
• Output: 50+ pages of structured research with sources
• Time: Minutes instead of hours

​

6. YouTube Analytics Channel

• API Access: Full YouTube statistics
• Queries: Natural language analytics questions
• Capabilities: Retention comparison, engagement analysis, topic performance
• Advantage: More flexible than YouTube Studio dashboards

​

7. Video Idea Research Channel

• Process: Continuous collection of links, articles, thoughts
• Enhancement: Agent enriches snippets and connects dots
• Timeline: Weeks of accumulated research ready for video creation
• Context: Builds understanding over time

​

8. Universal Summarization

• Input: Any URL (articles, videos, PDFs, research papers)
• Output: Structured summaries with key points
• Usage: Multiple times daily for content processing
• Integration: Can save directly to Obsidian

​

Infrastructure and Development

​

9. DevOps Management

• Migration: Automated ClawdBot to OpenClaw transition
• Maintenance: Killed zombie processes, fixed broken cron jobs
• Server Control: VPS management via API without SSH
• Security: Allow/deny lists for command permissions

​

10. Mobile Development

• Capability: Code from phone for quick fixes
• Scope: Bug fixes, simple features, pull requests
• Limitation: Not for production/main development workflow

​

Daily Life Management

​

11. Email Triage and Drafting

• Process: Reads inbox, flags important items, drafts responses
• Benefit: Same-day responses instead of weekend batches
• Security: Draft-only mode requires manual review

​

12. Calendar and Family Management

• Platform: WhatsApp group chat for family accessibility
• Function: “Schedule Dentist Thursday 3 PM” → automated booking
• Multi-user: Wife can add events, check schedules, get reminders

​

13. Voice Note Transcription

• Process: Whisper model transcribes voice messages to text
• Use Cases: Driving thoughts, shopping lists, meeting notes
• Integration: Can save important items to Obsidian or draft emails

​

14. Location-Based Services

• Coffee Shops: Google Places API with ratings, reviews, walking distance
• Shopping: Multi-store inventory checks (e.g., large snowboard boots)
• Intelligence: Comprehensive research in seconds

​

15. Weather and Reminders

• Advanced Forecasting: Warns of extreme conditions (e.g., -19°C cold snap)
• Smart Reminders: Rehab exercises with snooze, meeting alerts
• Context-Aware: More than basic weather apps

​

Social and Support

​

16. Technical Support for Friends

• Case Study: 3.5-hour OpenClaw setup assistance via WhatsApp
• Process: Screenshot debugging, step-by-step guidance
• Languages: Non-English support
• Outcome: 90% autonomous support, friend now uses own agent

​

17. Family Integration

• Group Chat: Wife + agent interactions
• Functions: Jokes, second opinions, conversation enhancement
• Real Impact: Agent reminded user to spend time with family

​

Advanced Architecture

​

18. Discord Migration Benefits

• Problem: Single conversation context pollution
• Solution: Dedicated channels per workflow
• Advantages:
  • Separated contexts
  • Per-channel model selection
  • Cost optimization
  • Better organization

​

19. Bookmark Management System

• Replaced: Paid Raindrop subscription
• Process: Drop link → summarize → extract key info → create tags
• Knowledge Graph: Connects related links over time
• Storage: Markdown in Obsidian, fully searchable

​

20. Obsidian Knowledge Base Integration

• Scale: 3,000+ notes indexed nightly
• Search: QMD semantic search (not keyword matching)
• Examples: “What did I decide about thumbnail design last month?”
• Rebuild: 10 seconds nightly after initial setup

​

Bonus Projects

​

WordPress Honeypot

• Purpose: Catch bots scanning for WordPress admin pages
• Implementation: Fake login page on non-WordPress site
• Security: Only on personal domain, temporary fun project

​

Excalidraw Integration

• Function: Automatic diagram and flowchart generation
• Use Cases: Architecture diagrams, concept maps, workflow visualization
• Efficiency: Minimal manual adjustments needed

​

Home Automation (In Progress)

• Platform: Home Assistant integration
• Control: Voice and chat-based smart home management
• Vision: Better than Siri for home automation

​

Starter Pack: Top 3 Use Cases

If overwhelmed, start with these three:

​

1. Draft-Only Email Triage

• Function: Catches missed items with urgent alerts
• Safety: Review before sending
• Impact: Immediate productivity boost

​

2. Daily Briefing to Markdown

• Function: Morning context, automatically organized
• Storage: Builds knowledge base over time
• Value: Structured start to each day

​

3. Discord Bookmark Channel

• Function: Drop links → automatic enrichment
• Replacement: Eliminates paid bookmark apps
• Growth: Knowledge base becomes more valuable over time

​

What Doesn’t Work Well

​

1. Memory Loss and Context Compaction

• Problem: Silent context compression loses important details
• Symptoms: Agent forgets mid-conversation, no warnings
• Mitigation:
  • Write everything to files
  • Use QMD semantic search
  • Manual compaction before automatic
  • Monitor context usage with status command

​

2. Cost Reality

• Issue: Opus is expensive but powerful
• Solution: Multi-model routing
  • Opus for deep thinking
  • Cheaper models for heartbeats/simple tasks
  • Discord channel architecture supports this
• Planning: Use cost calculator for budgeting

​

3. “What Do I Use It For?” Problem

• Reality: OpenClaw doesn’t invent workflows
• Requirement: Must have existing systems to automate
• Success Factor: People with established workflows see most value

​

4. Tasks Requiring Babysitting

• Complex Tasks: Multi-step processes still fail frequently
• Browser Automation: Flaky, sessions drop
• Reliability: Simple tasks = reliable, complex tasks = supervision needed
• Mitigation: Use sub-agents for complex tasks to preserve main context

​

5. Security Concerns

• Prompt Injection: No foolproof solution yet
• Email Content: Treat as potentially hostile
• Mitigations:
  • Draft-only email mode
  • Approval for destructive actions
  • Regular security audits
  • Tailscale for network security

​

Personal Failures and Lessons

​

Migration Issues

• Daily update cron job used old ClawdBot commands
• Failed silently for days after OpenClaw migration
• Missed updates due to package rename

​

Setup Complexity

• 3+ hours debugging authentication with friend
• Multiple false starts and reinstalls
• Setup intentionally difficult for security reasons

​

Context Management

• Hit compaction limits during complex research
• Now more aware and proactive about management
• Use sub-agents to preserve main context

​

Discord Migration

• Week of tweaking for right channel structure
• Finding optimal model assignments per channel
• Migrating context from Telegram conversations

​

Scoring Assessment

​

Setup Difficulty: 7/10

• Intentionally complex for security
• Requires technical knowledge
• Ensures user awareness of implications

​

Daily Value: 9/10

• Once properly configured and personalized
• Significant productivity improvements

​

Reliability:

• Simple Workflows: 8/10 - Very dependable
• Complex Tasks: 5/10 - Still needs improvement

​

Best Features

• Discord channel architecture with per-channel models
• File-based memory with markdown-first approach
• Nightly semantic indexing builds valuable knowledge base

​

Most Useful

• Background heartbeat checks - quietly catches important items
• Growing intelligence - learns preferences over time

​

Biggest Pain Point

• Memory and context compaction - still needs work

​

Surprises After 50 Days

​

Positive Surprises

• Gets better over time - learns user preferences and style
• Anticipates needs - proactive suggestions based on context
• Replaced more than expected: Zapier, Raindrop, YouTube Studio analytics, Apple Shortcuts
• Cost savings: No longer paying for multiple SaaS tools

​

Learning Curve

• Week 1: Simple tool
• Week 3: Feels like infrastructure
• Week 7: Reorganize workflows around it
• Transforms from chatbot to system

​

Ecosystem Growth

• Thousands of skills on ClawHub
• Hosted services for non-technical users
• Rapidly maturing tooling and security
• Community growing from hundreds to tens of thousands

​

Recommendations

​

Yes, Recommend If:

• You have workflows to automate
• Comfortable with terminal/technical setup
• Understand cost implications
• Enjoy building and iterating systems

​

Not Yet Recommended If:

• Want out-of-box solution
• Not technical
• Expect fully autonomous operation
• Need guaranteed reliability for critical tasks

​

Future Outlook

• Currently using ~5% of potential capability
• Extremely high ceiling for capabilities
• Floor still has reliability holes
• Security and setup will improve over time
• Eventually everyone will run some version of this
• Most fun with technology in years

​

Key Resources

All prompts and use cases available in provided GitHub document for copy-paste implementation. Includes:

• Complete prompt library
• Security tips
• Cost optimization strategies
• Setup documentation
• Community directory (Clawdiverse.com)

​

Final Assessment

After 50+ days of daily use through multiple iterations and rebrands, OpenClaw represents a fundamental shift from chatbot to system. While challenges exist around memory management, cost optimization, and setup complexity, the productivity gains and workflow automation potential make it worthwhile for technical users with existing systems to enhance. The strongest endorsement: “I’m not going back.”