-
-
Bangla homepage
-
Complaint intake
-
Evidence review
-
Prepared draft
Inspiration
When an online purchase goes wrong, the receipt, payment proof, seller messages, and delivery details can be scattered across different apps. I wanted to make it easier for people in Bangladesh to organize that information and understand what they might do next.
What it does
KenaSathi helps users record purchase details, organize evidence, build a case timeline, and prepare a complaint draft they can review. It includes Bangla and English guidance, FAQs, and optional AI assistance for summaries and fact suggestions. Users remain in control: KenaSathi does not submit complaints or guarantee an outcome.
How we built it
I built the web app with React and TypeScript, with a Node.js help service using SQLite for FAQ and support features. Case details and evidence are stored in the user’s browser. Optional AI assistance uses a provider only when the user supplies a temporary API key and gives consent. The app also uses local rules for date calculations and guidance.
Challenges we ran into
One challenge was handling dates carefully. The relevant complaint period may depend on when the cause of action arose, which can differ from the purchase date. I designed the guidance to avoid guessing when that date is unclear.
Privacy and trust were also important. Evidence may contain sensitive details, so users need to know where it is stored and when information may be sent to an AI provider. I also worked to make the interface useful in both Bangla and English while clearly explaining that KenaSathi is an independent tool, not a government service.
Accomplishments that we're proud of
I’m proud to have a working public app that brings purchase records, evidence, timelines, and complaint-draft preparation into one place. I’m also proud that the product keeps users in control, provides a fictional sample for exploring the workflow, and avoids promising complaint acceptance, refunds, or rewards.
What we learned
Building this taught me that AI is most useful when people can inspect and correct its suggestions. It also showed me that trustworthy consumer tools need clear privacy explanations, careful data handling, and plain language—not just more features.
What's next for KenaSathi: Consumer Rights & Complaint Draft Assistant
Next, I’d like to test the app with real Bangladeshi consumers, improve the mobile and accessibility experience, and make the AI demonstration easier for judges to try without exposing a private API key. I’d also like to keep reviewing the guidance and support information against official sources.
Built With
- accessibility
- artificial-intelligence
- bangla
- consumer-rights
- google-gemini-api
- node.js
- openai-api
- pdf.js
- privacy
- react
- sqlite
- tailwind-css
- typescript
- vite
- zod
