Over the past few months, Toket has been trying to answer one question: Can an independent Builder create a truly valuable AI product? At the beginning, I simply wanted to build small tools that AI users might need.
AI Cost Analysis. Prompt Optimization. Model Information. Workspace. GEO. Every feature was an experiment. I wanted to understand: In the AI era, can one person use modern tools, models, and development workflows to build a complete product?
But as Toket grew, another question became more important: If someone who knows nothing about Toket visits for the first time, will they understand what it is?
⸻ In the early stage, the English website was mainly an international version of the Chinese website. Pages could switch languages. Content could be translated.
Features could be used. But the real challenge was not language. It was product communication. Different markets understand products differently.
A user visiting from overseas needs clearer answers: What is Toket? What capabilities does it provide? Why do these capabilities matter? So this update focused on rebuilding ToketAI.com.
The English website is no longer just a translated version. It is becoming an independent product entrance. ⸻ One of the biggest changes was GEO.
A few months ago, GEO was mainly an exploration capability. I wanted to understand: Can AI systems understand brands? Can user questions trigger brand recommendations?
Can websites be understood by AI? But as the testing and reporting system matured, I realized: GEO should not only be an internal analysis tool.
It can become a real product. So this update created a clearer GEO product path. Free: Free Website AI Readiness Check. Answering: Can AI access and understand my website?
Paid: AI Brand Visibility Audit. Answering: How does AI understand, describe, and recommend my brand? Professional: GEO Optimization Service.
Providing: AI visibility optimization. Content and brand structure improvements. Future validation and re-testing. For the first time, GEO moved from technical capability into a complete product system.
⸻ Another important change was the GEO reporting system. Previously, GEO testing was mostly my own research. I ran tests. Analyzed results. Documented problems.
But if it becomes a real service, it needs to support real customers. Customers may come from different regions. They may need different languages.
So this update introduced English report delivery. GEO reports can now support: Chinese versions. English versions. Customer-facing versions.
The report structure remains consistent: Executive Summary. AI Brand Visibility. Key Breakpoints. Priority Actions. Evidence & Appendix. The original AI outputs remain in their source language.
Because real evidence matters more than translated text. ⸻ During this internationalization process, one thing became especially clear: Removing Chinese-market leftovers.
Many products fail at internationalization not because translation is poor. They fail because the product never truly changes markets. Examples:
Showing RMB pricing. Using unverified contact emails. Switching languages after page load. Displaying Chinese placeholders. These details directly affect user trust.
So this update was not only translation. It was reviewing every detail a global user experiences. ⸻ At the same time, I reorganized Toket’s public structure.
The current navigation is clearer: Tools: AI Cost Analysis. Prompt Optimizer. GEO. Knowledge: AI News. AI Models. Laboratory: Toket Lab Notes.
About: About Toket. Each section now has a clear responsibility. Tools solve problems. Knowledge builds assets. Laboratory records the process.
About explains identity. ⸻ During this development cycle, something interesting happened again. Cursor kept turning development into a series of choices.
Previously, AI Coding was mostly about: “How do I implement this?” Now the questions are increasingly: “Which direction should I choose?” Should this page stay or disappear?
Should this feature expand or become less visible? Should this capability become commercial or remain experimental? These are no longer coding problems.
They are product decisions. AI can help me explore options faster. But the final product direction still requires the Builder. ⸻ This update did not add many new features.
Most of the work was reorganizing existing capabilities. But I believe this is an important step in Toket’s transition from experiment to product.
Because a real product needs more than features. It needs: A clear identity. A trustworthy message. A stable experience. And a value path users can understand.
Toket is still small. There is no large team. There is no huge marketing budget. But this is the first time I am no longer only testing an AI product for myself.
I am trying to make Toket understandable, usable, and trustworthy for someone who has never heard of it before.
Estimate task cost in the AI Cost Analysis or refine prompts in the Prompt Optimizer.