Build Journal
SEO Improvements & Humanization in Bundles — September 29, 2026
I tackled SEO adjustments and enhanced content humanization in bundles today, improving our platform's performance and user experience.
What shipped
- SEO Case Sensitivity Fix — Redirected 301 case variants to lowercase canonical URLs.
- Content Humanization Update — Adjusted article length for better humanization from 300 to 250 words.
- Whole Page Humanization — Implemented full page humanization for articles and FAQs.
- Humanization Recipe Restoration — Restored the humanization process to effectively pass AI detection.
- Bulk Bundle Rewrite Optimization — Enhanced bulk bundle rewrites through shared humanizer.
Today, I focused on refining our SEO strategy and enhancing the humanization process for content bundles. After a long 16-hour day filled with coding and troubleshooting, I managed to ship 15 commits, including significant improvements and bug fixes. The goal was to ensure that our platform is not only more efficient in handling SEO but also capable of producing more engaging and human-like content.
One of the standout features I implemented was the adjustment of 301 redirect rules for our URLs. Previously, case sensitivity in our routing led to some published URLs redirecting incorrectly instead of producing a 404 error. This was a critical issue because it affected how search engines indexed our pages. By ensuring that all case variants of our hashtags and local URLs are redirected to a lowercase canonical version, I can streamline our SEO efforts and eliminate potential indexing issues.
On the content front, I redefined how we humanize our articles and FAQs. The previous threshold for article length was set at 300 words, but I found that this often skipped over 90% of real pages, leaving us with a lot of thin content. I lowered the limit to 250 words, which allows us to humanize the majority of our articles while still filtering out the less valuable stubs. This change was crucial as it aligns better with our median article length of around 292 words, thus improving the quality of our content.
In addition to adjusting word counts, I also worked on humanizing the entire page, which includes both the article and every FAQ answer. This was a significant technical challenge, especially considering the need to ensure that thin pages do not ship un-humanized. By gating this process on a full ~350-word article, we can ensure that only quality content makes it through while maintaining a balance between user experience and SEO performance.
Another major undertaking was restoring the humanization recipe that effectively beats AI detection systems like GPTZero. I had to revert to a single-pass method that I had previously tested. This involved changes in our humanizer’s configuration to ensure every paragraph was rewritten, which is essential for passing AI detection. The aggressive and bursty rubric I implemented guarantees that we have a clean rewrite score of approximately 0.999 against detection systems, ensuring that our content remains undetected as AI-generated.
I also made significant revisions to our bulk bundle rewrite processes, routing them through our shared humanizer to leverage the same data gates as our API. This allows us to prioritize interactive calls, improving our response time and overall user satisfaction. I implemented several scripts to regenerate and rehume samples efficiently, a necessary step as we scale.
Throughout the day, I encountered a few bugs that ate into my time. One was related to our ticker feature, which required moving the live balance pill to a more user-friendly position on the page. This was a minor fix, but it took longer than anticipated to ensure that it did not disrupt the overall layout. Additionally, I had to halt the auto-deploy from GitHub Actions as it was causing race conditions post-migration, which led to failed deployments. Switching to a more controlled on-box deployment method allowed me to regain stability in our release process.
As I reflect on today’s accomplishments, I’m reminded of the unique challenges of building this platform solo, with AI as my only collaborator. I often think about how tools like Claude have transformed my workflow, allowing me to handle complex tasks that would otherwise require a larger team. This is particularly evident in the humanization improvements where Claude’s assistance has been invaluable. It’s moments like these that reinforce my ambition to build a one-man-show company aiming for a billion-dollar valuation. Each feature I ship not only enhances our platform but also brings me one step closer to that goal.
As I wrap up this long day, I'm excited about the progress made. The improvements in SEO and content humanization are critical for the success of our platform, and I’m eager to see how users respond to these changes. I know that the road ahead will be filled with more challenges, but with each commit, I feel more equipped to tackle whatever comes next. Today was a clear reminder that while the grind is real, the rewards of building something impactful make it all worthwhile.