Local Fetch Improvements & Tavus Unblocking for Agentic Web — September 4, 2026

Today, I set out to enhance our local data fetching capabilities and address some persistent issues with the Tavus integration. The goal was to ensure smoother interactions within the agentic web environment, particularly for users engaging with our radio features and the Tavus video personas. I shipped one new feature, improved local fetch capabilities, and resolved two bugs related to Tavus and radio functionalities.

In the first commit, I implemented the untrackedFetch functionality for external calls, which is a significant upgrade for rendering paths in our local module. This change allows the portal CRM to fetch data more efficiently without getting bogged down in tracking every single request. The ability to purge noindex tags as part of this process is a game changer. I spent about an hour focused on this, and while it felt like a small step, it lays the groundwork for more robust local SEO and user engagement.

Next, I turned my attention to fixing the Tavus integration. The second commit was all about unblocking the in-app frames, specifically for the AISEO board and Calendly booking tools. These frames were essential for providing a seamless user experience, but they were encountering issues that prevented them from displaying properly within the application. I traced the problem back to the interaction configuration for the radio-stage tools, which required grounding. After some debugging, I managed to resolve the issue and ensure that users can now access these tools without any hitches. It felt great to see it all come together, but it also reminded me of how delicate these integrations can be.

The final commit was a fix for the radio functionality, where I corrected the implementation to ensure that Lucas live uses the OWNED Interviewer persona instead of the tool-less stock ID. This was critical because the persona adds a layer of authenticity and engagement that users expect when tuning into our radio features. Diagnosing this bug took longer than I anticipated, and it was a bit frustrating to trace back through the code to identify the root cause. But ultimately, it was worth it. The interaction feels much more genuine now, and I can already see how this will enhance user experience.

Reflecting on the day, it’s fascinating to see how each of these components plays into the broader vision of building a one-man-show company with a billion-dollar valuation. I’m doing this solo, and while it can be daunting, I’m leveraging AI tools to support my development process. For instance, using Claude Code in VS Code has been instrumental in speeding up my workflow and allowing me to tackle these features and fixes efficiently. I’m not just building an agentic web platform; I’m crafting a robust ecosystem that will enable dynamic interactions across various tools and services.

As I continue to refine the product, I’m learning valuable lessons about prioritization and focus. Every hour spent on these improvements adds up, and I can see the potential for exponential growth as I streamline these processes. The combination of local fetch capabilities and enhanced Tavus integration is a step toward creating a more cohesive experience for users navigating the agentic web, and I’m excited to see where it leads.

I’m looking forward to the next steps and tackling the challenges ahead. Each day brings new insights and opportunities for improvement, and I’m committed to pushing the boundaries of what’s possible. The agentic web is more than just a platform; it’s an evolving journey, and I’m here for every twist and turn.