I just replaced myself with Clawdbot… here's how

Claudebot: Setup, Use Cases, and Deep Dive into a Self-Hosted AI Assistant

Key Concepts:

• Claudebot: A self-hosted AI assistant capable of interacting with various applications via chat interfaces (WhatsApp, Telegram, etc.).
• AGI (Artificial General Intelligence): The hypothetical intelligence of a machine that could understand, learn, adapt, and implement knowledge like a human.
• VPS (Virtual Private Server): A virtual machine sold as a service, offering dedicated resources on a shared physical server.
• OpenRouter: A platform providing access to multiple AI models through a unified API.
• 11 Labs: An AI voice generation platform.
• Codex/Claude: Large language models developed by OpenAI and Anthropic respectively.
• Ollama/LM Studio: Tools for running large language models locally.
• TUI (Terminal User Interface): A text-based user interface for interacting with software.
• Bun/npm: Package managers for JavaScript.
• OpenAI Whisper: An automatic speech recognition (ASR) system.

I. Introduction & Capabilities

Claudebot is presented as a powerful, self-hosted AI assistant gaining popularity, even surpassing Cloud Code in usage. It distinguishes itself by its ability to “actually do things” – managing tasks like clearing inboxes, sending emails, managing calendars, and checking in for flights – directly within existing chat applications like WhatsApp, Telegram, and others. A key benefit is data privacy, as it runs on the user’s own machine, avoiding reliance on companies like Anthropic or OpenAI for data handling. It boasts persistent memory, browser control, full system access, and extensibility through skills and plugins. The presenter highlights a demonstration by Alex Finn where Claudebot, upon failing to secure a restaurant reservation through OpenTable, autonomously used 11 Labs to generate a voice call and complete the booking, framing this as a potential indicator of AGI arriving. Another example showcases Aaron using Claudebot to manage multiple AI agents (Codex and Claude) for feature development, including code review and PR merging.

II. Setting Up Claudebot: A Step-by-Step Guide

The most affordable setup method involves using a VPS rather than purchasing expensive hardware like a Mac Studio. The presenter demonstrates the installation process using Hostinger as the VPS provider.

1. Accessing the Terminal: Log into the Hostinger panel and access the terminal (SSH).
2. Running the Installation Command: Copy and paste the “oneliner” command from the Claudebot website (cloth.bot) into the terminal. This command automatically installs NodeJS, Git, and Cloudbot.
3. Security Warning: Acknowledge the security warning, recognizing the agent’s powerful system access. Running on a VPS is strongly recommended to mitigate risks.
4. Onboarding & Quick Start: Select “quick start” during onboarding.
5. Model Provider Selection: Choose a model provider (OpenAI, Gemini, or OpenRouter). The presenter recommends OpenRouter for access to a wider range of models.
6. API Key Configuration: Obtain an API key from OpenRouter (openrouter.ai), create an account, and paste it into the terminal. The presenter emphasizes the importance of keeping API keys secure.
7. Channel Selection: Choose the desired chat application (WhatsApp, Telegram, Discord, etc.). The presenter selects WhatsApp.
8. WhatsApp Linking: Scan the QR code generated by Cloudbot using WhatsApp on a smartphone to link the accounts.
9. Skill Installation: Select desired skills (Apple Notes, Apple Reminders, Gemini, Google Places, Obsidian, etc.). The presenter skips this step for a future dedicated video.
10. API Key for Whisper: Obtain an OpenAI API key from the OpenAI API platform (platform.openai.com) and paste it into the terminal to enable speech-to-text functionality.
11. Finalization: Complete the onboarding process and verify functionality.

III. Advanced Use Cases & Functionality

The video details several impressive use cases for Claudebot:

• Dan’s Workflow: Time blocking tasks in a calendar based on importance, weekly reviews based on meeting transcripts and notes, and notifying family members about school tests.
• Research & Briefing: Breaking down large projects into smaller tasks, researching individuals before meetings, and creating briefing documents.
• Autonomous Idea Validation: Spawning sub-agents to research business ideas.
• Calendar Management: Resolving calendar conflicts autonomously.
• Invoicing & Summarization: Creating invoices and summarizing work.

These examples illustrate Claudebot’s potential to act as a personal assistant, automating complex tasks and improving productivity.

IV. Debugging & Troubleshooting (Real-Time Demonstration)

The presenter encountered issues during the setup process, specifically with model recognition and communication. He demonstrates a real-time debugging session using OpenCode, a code editor installed on the VPS.

• The Problem: Claudebot was failing to recognize the selected model (Opus 4.5) and was stuck in a loop.
• The Solution: Using OpenCode, the presenter identified a minor error in the configuration file (a missing dot) and corrected it. This involved running commands directly on the VPS and analyzing the system logs.
• Key Takeaway: The presenter emphasizes the importance of being comfortable with the terminal and basic debugging techniques for effectively managing a self-hosted AI assistant.

V. Hostinger Sponsorship & VPS Benefits

The video includes a sponsored segment promoting Hostinger as an affordable and easy-to-use VPS provider. The presenter recommends the KVM2 plan (starting at $7/month) with 2 vCPU cores, 8 GB of RAM, 100 GB of disk space, and 8 TB of bandwidth. Using a coupon code "David" provides an additional 10% discount. The benefits of using a VPS are highlighted: affordability, data privacy, and avoiding the need for expensive hardware.

VI. Final Remarks & Future Content

The presenter concludes by reiterating the power of Claudebot and encouraging viewers to explore its capabilities. He plans to create a follow-up video detailing advanced features like integration with 11 Labs, Nadubana Pro, OpenAI Whisper, and Google Calendar. He encourages viewers to subscribe to receive notifications about future content.

Notable Quotes:

• “AGI is here and 99% of people have no clue.” – Regarding the restaurant reservation example.
• “This might be the single most insane use case I’ve seen of CloudB showing its power.” – Referring to Alex Finn’s demonstration.
• “The setup is the hardest part. Once you set it up, if you can sit down for a couple minutes and actually push through it, you'll get all of the benefits.” – Emphasizing the initial learning curve.

Data & Statistics:

• Hostinger KVM2 plan: $7/month, 2 vCPU cores, 8 GB RAM, 100 GB disk space, 8 TB bandwidth.
• Mac Mini price: $600+
• VPS cost: ~$7/month

Logical Connections:

The video follows a logical progression: introduction of Claudebot, demonstration of its capabilities, step-by-step setup guide, real-time debugging session, and concluding remarks. The troubleshooting segment highlights the practical challenges of self-hosting and the importance of technical skills. The Hostinger sponsorship is seamlessly integrated, providing a solution to the hardware cost barrier.

Synthesis/Conclusion:

Claudebot represents a significant step towards accessible and powerful AI assistance. Its self-hosted nature, combined with its ability to integrate with existing chat applications, makes it a compelling alternative to cloud-based solutions. While the setup process requires some technical proficiency, the potential benefits – increased productivity, data privacy, and customization – are substantial. The presenter’s demonstration and debugging session provide valuable insights into the practical aspects of deploying and managing this innovative AI assistant. The video serves as a strong call to action for those interested in exploring the future of AI and taking control of their own data.