JapaDesk

The process of applying for graduate programs abroad (colloquially known as "Japa" in Nigeria) is a grueling, fragmented logistical nightmare. Prospective students track university deadlines on Excel, draft essays in Word, calculate complex credential conversions manually, and scour Reddit for visa requirements. Missing a single deadline or misunderstanding a specific university's GPA conversion rule can cost a year of progress.

While there are many fragmented tools and forums, there was no single "Operating System" tailored to the entire lifecycle of studying abroad—from shortlisting universities to tracking applications and finally securing a visa. Furthermore, generic AI chatbots hallucinate too much for critical admission tasks. I realized I could use closed-context AI to provide specific, reliable assistance.

I designed JapaDesk as a centralized hub. The core organizing principle is a Kanban board for application tracking, enriched with automated countdown timers. Around this core, I built a suite of specialized tools handling the most painful parts of the process: grading conversions, document preparation, and financial planning. The AI integration had to be strictly grounded to prevent harmful hallucinations regarding university requirements.

The platform features a comprehensive Kanban board, an AI Syllabus Fit Check, and an SOP (Statement of Purpose) coach that evaluates drafts against multiple rubrics. I built a robust international credential converter that accurately handles the German Bavarian Formula, US 4.0 GPA, UK Honours, and French 20-point systems. To manage the relocation phase, it includes a 16-country visa checklist suite and a Proof of Funds simulator tailored for systems like the German Sperrkonto, UK 28-day rule, and Canadian GIC. Finally, I added a CRM for managing professor outreach and recommender coordination.

I chose Next.js and Supabase for a reliable, real-time foundation. For the AI features, I integrated the Anthropic Claude API. I specifically chose Claude for its larger context window and superior ability to adhere to strict system prompts, which is crucial for the closed-context syllabus fit checks. By grounding the AI queries in provided university documentation, the system acts as a reliable assistant rather than a generic text generator.

Building the credential converter was surprisingly complex. The rules for converting a Nigerian 5.0 CGPA to the German Bavarian formula aren't strictly linear and require handling various edge cases and grading scales across different universities. Implementing the closed-context AI prompts required extensive testing and tweaking to ensure the model wouldn't confidently output incorrect university requirements. Ensuring the UI remained uncluttered despite the vast amount of tools and data was a continuous design challenge.

JapaDesk successfully centralizes a chaotic process into a manageable workflow. The feedback on the Proof of Funds simulator and the credential converter has been particularly strong, as these tools save users hours of manual calculation and anxiety. It acts as a comprehensive digital companion for a very stressful life transition.

Looking back, I would have integrated a community or peer-review aspect earlier. While the AI SOP coaching is powerful, getting feedback from successfully admitted students is invaluable. Future iterations will likely include a way to securely share application profiles with mentors or peers for human feedback.