Tadah

Inspiration

In a lot of families, one person ends up reading everybody’s mail.

“What does this say? Is it bad? Can you deal with it?”

This happened in my own family. But translating a letter is only the beginning. There’s still a deadline to remember, a form to figure out, or a phone call to make in English.

I built Tadah for families like mine—and anyone who’s ever opened a letter, felt overwhelmed, and kept putting it off.

What it does

Tadah is an iPhone app that turns confusing paperwork into clear next steps.

Scan a letter or bill, and Tadah explains it in English, Spanish, or Chinese. It identifies important details and deadlines, creates a checklist, and lets you translate the full document.

You can organize tasks by family member, set reminders, and follow guides for everyday paperwork.

If the next step is a phone call, Tadah can make it for you. You review the recipient, the question, and the permissions first. The phone agent identifies itself as AI, calls in English, and brings the answer back in your language.

How we built it

I built Tadah with Swift and SwiftUI, using Apple Vision to read scanned text and OpenAI models to help explain documents and generate next steps.

SwiftData stores tasks locally, so users can access saved information offline. Cloudflare Workers handles backend requests and account sync, with Cloudflare D1 storing account and task data.

CALL-E powers the phone agent, and RevenueCat manages subscriptions. I used Claude and Codex to help write and review code throughout development.

Challenges we ran into

Documents often contain several amounts and dates. Reading them correctly doesn’t automatically tell you what someone owes or when they need to act. I combined parsing with AI interpretation and made the extracted information editable before saving.

Calling introduced another challenge: a request can time out even when the phone is already ringing. I added request tracking and recovery logic so an uncertain response doesn’t automatically cause a duplicate call.

Keeping explanations, instructions, and call results in the user’s chosen language required attention throughout the app. So did syncing changes across devices without mixing up accounts or bringing back deleted tasks.

Accomplishments that we're proud of

I shipped Tadah to the App Store and connected the full journey: scanning a document, understanding it, taking action, and remembering what’s left.

I’m especially proud of the calling flow. “Call this number” can be the hardest item on someone’s checklist. Tadah gives users a way to get help with that step while reviewing the request before the call begins.

Seeing a problem from my own family become a working app is the part that means the most to me.

What we learned

Understanding a letter and handling it are different tasks. A useful explanation still needs a clear next step—and sometimes help carrying it out.

I also learned that a completed call doesn’t necessarily mean a completed task. Someone may still need to pay a bill, send a document, or wait for a response. The app needs to make that remaining work clear.

Building with AI taught me to review and test carefully, especially when a feature can take action outside the app.

What's next for Tadah

I want to learn where people still get stuck after scanning their first letter, and whether Tadah helps when the next one arrives.

That feedback will guide improvements to explanations, guides, and calling. Longer term, I’d like Tadah to help with more follow-up actions, such as filling out forms, with users reviewing important details before anything is submitted.

The goal stays simple: help people handle everyday paperwork with more confidence and less dependence on someone else.

Built With

• apple
• call-e
• claude
• cloudflare
• d1
• javascript
• openai
• revenuecat
• swift
• swiftdata
• swiftui
• vision
• workers
• xcode