The video reviews Clawbot (now Moldbot), an open-source AI assistant that runs 24/7 on your device, automates tasks across apps, and is highly customizable, but requires technical setup and careful attention to security and privacy risks. While powerful and community-supported, its features overlap with existing automation tools, making it most valuable for tech-savvy users seeking persistent, flexible AI automation.
The video discusses the recent viral rise of Clawbot (now renamed Moldbot), an open-source AI agent that stands out because it can run 24/7 on your device or server, remember context, and access your files and apps if you grant permission. Unlike typical browser-based chatbots, Clawbot can be reached via messaging platforms like Telegram or WhatsApp and can automate a wide range of tasks, from managing emails and calendars to controlling smart home devices. The code is freely available on GitHub, and users can connect it to various AI model providers, making it highly customizable.
The creator debunks several myths and overhyped claims circulating online, such as the false idea that you need a Mac Mini to run Clawbot. In reality, it works best on hardware with an Nvidia CUDA GPU, but it can also be set up on a free AWS server. The video provides a step-by-step installation guide, including how to launch an AWS EC2 instance, install Clawbot, and connect it to an AI model provider like Zhipu AI (ZAI), which is highlighted as a cost-effective alternative to Anthropic’s Claude.
Practical use cases for Clawbot include automating email summaries and responses, scheduling meetings, trading prediction markets, managing social media posts, and controlling smart home devices. The video demonstrates how to connect Clawbot to Telegram, install additional skills (such as reminders), and enable web search using the Brave Search API. The process involves several technical steps, including obtaining API keys and configuring integrations, which the creator notes are often glossed over by other influencers.
A significant portion of the video is dedicated to security and privacy concerns. The creator warns that linking Clawbot to sensitive accounts like Gmail or Google Drive can expose users to risks such as prompt injection attacks or accidental data loss. Running Clawbot locally without proper sandboxing could give the AI access to all your files, and using external AI providers means your data is sent to their servers, which is not private. The video recommends running regular security audits and, for maximum privacy, using local AI models via Ollama.
Finally, the creator puts Clawbot’s capabilities in perspective, noting that while it is powerful, open-source, and supported by a strong community, many of its features already exist in other automation tools like n8n, GenSpark, and Manis. The video concludes that Clawbot is useful for automating repetitive tasks and acting as a persistent AI assistant, but it requires technical know-how and careful attention to security. Viewers are encouraged to share their experiences and subscribe for more AI news and tutorials.