The video demonstrates using the open-source GLM 5.2 model combined with an autonomous agentic harness via the Open Code app to perform a comprehensive, cost-effective AEO audit and generate professional lead materials without expensive subscriptions or frontier AI models. It highlights the flexibility, affordability, and sovereignty of this multi-agent AI approach for various applications, encouraging viewers to adopt and customize these tools amid growing restrictions on AI access.
In this video, the creator demonstrates running GLM 5.2 through an autonomous, agentic harness using the Open Code desktop app to perform a comprehensive AEO (AI, SEO, and web audit) on a prospect’s website. The harness not only conducts the audit but also generates a cover letter and executive summary to use as a lead magnet for pitching services. The key takeaway is that this setup operates without requiring expensive subscriptions or frontier AI models, making it a cost-effective and sovereign AI solution amid increasing geopolitical restrictions on AI access.
The video begins by showcasing a lead generation harness that autonomously scrapes and processes leads from various sources using Apify, costing under $5 and running for about 45 minutes without human intervention. This highlights the power of open-source AI models like GLM 5.2 combined with agentic harnesses to perform complex tasks efficiently and affordably. The creator emphasizes the importance of maintaining control over AI tools as access to frontier models becomes more restricted and costly.
Next, the creator walks through running a full AEO audit harness found on GitHub, which involves multiple specialized agents working together to crawl a website, analyze its HTML, assess SEO and AEO metrics, research competitors, and generate a detailed report. The harness also includes quality control agents that review and revise the output to ensure accuracy and professionalism. This multi-agent approach mimics a team of experts collaborating, which improves the quality of the audit compared to a single AI model working alone.
The video also covers how to build custom harnesses by adapting existing templates, such as a screenplay-writing rig or an academic thesis writer, demonstrating the flexibility of this approach for various use cases. The creator shows how to configure and run these rigs using GLM 5.2, highlighting the cost savings compared to using more expensive models like Opus or Sonnet. The Open Code app is praised for its user-friendly interface, although it has some limitations, such as file visibility and occasional interruptions.
Finally, the creator concludes by sharing the results of the AEO audit, including a detailed report and an executive summary that can be sent to prospects as a lead magnet. The entire process took about an hour and cost roughly $9 in API usage, significantly cheaper than commercial alternatives. The video encourages viewers to learn how to build and use these open-source AI harnesses to maintain sovereignty over their AI capabilities in an uncertain future, inviting them to join the AI Captain’s Academy for further learning and community support.