Cloudbot Practical Usage Guide

Overview

ClowdBot, now referred to as Moldspot in parts of the ecosystem, is not a mobile AI app or a chat interface.

ClowdBot is a persistent AI agent running on a real Mac machine with:

• Screen access
• Mouse and keyboard control
• Memory
• Autonomy through background execution

You message it like a coworker.

It performs tasks like a human sitting at that computer.

System Requirements and Setup Philosophy

ClowdBot can run on:

• A Mac Mini
• A Mac Studio
• A rented virtual Mac

The transcript strongly recommends simplicity over complexity.

Beginner recommendation:

• Use the cheapest Mac Mini available
• Or rent a virtual Mac
• Avoid VPS and Docker debugging

Reason:

You need visual access to the screen so you can step in when ClowdBot gets stuck.

Core Setup Process Summary

The setup requires only two skills:

1. Copy and paste
2. Opening Terminal

High level setup flow:

1. Install Homebrew
2. Install NodeJS and npm
3. Install ClowdBot
4. Select Anthropic connection
5. Connect Telegram
6. Skip advanced skills initially
7. Enable required hooks
8. Verify ClowdBot responds to messages

Once ClowdBot replies with a greeting, setup is complete.

Use Case 1: Proactive Daily Accountability

What this does

ClowdBot initiates conversations on a schedule.

You do not need to prompt it.

How it is implemented

ClowdBot uses cron jobs to send messages at fixed times.

Example workflow from the transcript:

• Morning message asks for top 3 priorities
• Evening message asks what was completed
• Asks why items were not completed
• Logs everything automatically
• Syncs to ClickUp

How the user interacts

The user replies with:

• Text
• Voice notes

ClowdBot handles transcription and logging.

Outcome

This system replaces self discipline with structure.

Accountability becomes automatic.

Use Case 2: Voice Notes and Transcription

What this does

ClowdBot accepts voice messages and converts them into actionable data.

Setup steps from the transcript

1. Open Skills section
2. Filter for transcription
3. Add OpenAI API key
4. Save key

How it is used

• Send a voice note via Telegram
• ClowdBot transcribes it
• Converts it into tasks or logs

Outcome

Tasks are captured while:

• Walking
• Traveling
• Away from keyboard

No typing required.

Use Case 3: Autonomous Browser Control

What this does

ClowdBot controls a real browser on the Mac.

Example shown in transcript

The user asks ClowdBot to:

• Go to YouTube
• Find latest video
• Report view counts

ClowdBot:

• Opens the browser
• Navigates on its own
• Tries multiple approaches if something fails
• Reports results in Telegram

Setup requirement

• Install ClowdBot Chrome extension
• Enable developer mode in Chrome
• Load extension manually

Outcome

ClowdBot performs research without supervision.

Use Case 4: Keeping ClowdBot Running 24 by 7

Problem addressed

Mac machines sleep.

Sleeping machines stop agents.

Solution shown

1. Disable automatic sleep
2. Enable restart after power failure
3. Run terminal keep awake command
4. Verify caffeinate is running

Result

ClowdBot stays active continuously.

Use Case 5: ClowdBot as Project Manager

What this does

ClowdBot becomes a task tracking layer.

How it works

• ClowdBot has its own email
• Tasks are logged into ClickUp
• Status updates automatically
• Multiple tasks run in parallel

The user does not manually update task states.

Key behavior

ClowdBot logs work while doing the work.

Use Case 6: Sponsorship and Email Negotiation Agent

Problem solved

Sponsorship emails cause time drain.

Solution shown

1. Create an AI email inbox using AgentMail
2. Train ClowdBot on:
   • Rates
   • Non negotiables
   • Approval rules

Workflow

• User forwards sponsorship email
• ClowdBot drafts reply
• Requests approval
• Sends only after confirmation

Result

Negotiation scales without manual back and forth.

Use Case 7: Automated Marketing and Content Distribution

What this does

ClowdBot acts as a marketing team.

Workflow described

• New YouTube video published
• Trigger fires automatically
• Video sent to clipping service
• Shorts generated
• Scheduled across platforms

Tools involved

• Opus Clip
• Blotato
• Automation scripts written by ClowdBot

The user does not write the code.

Use Case 8: Website QA and Automated Fixes

What this does

ClowdBot audits websites.

Example shown

• Checks footer links
• Validates redirects
• Identifies broken pages

Advanced capability

ClowdBot has access to GitHub repository.

It can:

• Modify code
• Fix issues
• Push changes

All triggered by a single instruction.

Remote Access and Mobility

Remote desktop setup

• Jump Desktop installed
• One time payment
• No subscription

This allows:

• Phone access
• Tablet access
• Remote monitoring

ClowdBot becomes a portable server.

Key Mental Model from the Transcript

ClowdBot is not prompt driven.

ClowdBot is intent driven.

You:

• Describe goals
• Stay intentionally vague
• Let ClowdBot figure execution

You monitor results, not steps.

Final Summary

ClowdBot functions as:

• A proactive assistant
• A browser operator
• A project manager
• A marketer
• A QA engineer
• A negotiator
• A persistent automation server

The value is not intelligence.

The value is persistence and autonomy.

Once set up, ClowdBot works while you do something else.