Amanah App : Charity and Funding Report System

Accurate and transparent Reporting is the Lifeblood of Charity. To ensures that every contribution achieves its intended impact, fostering a sustainable cycle of generosity and hope.

Comment

## Inspiration

Sumatera Disaster 2025 is shocking the world. Not only Sumatera, there are so many disasters happen in the world. The victims need help. And The donors need transparent and reliable report. That is why I develop Amanah Report App. The word Amanah translates to "Trust" or "Accountability." In the world of charity, donors often give with their hearts but remain disconnected from the actual impact. NGOs, on the other hand, are often overwhelmed by fieldwork and don't have the time to produce professional, transparent reports. We were inspired to bridge this gap using AI—turning raw, messy field data into a standardized system of radical transparency.

## What it does

Amanah Report is an end-to-end audit and reporting ecosystem.

  • Field Input: Staff can upload photos of food distributions, record voice notes about beneficiary stories, and scan physical receipts/bills.
  • AI Analysis: The system uses Gemini 3 Flash to instantly extract line items from bills, transcribe local dialects in voice notes, and analyze images to verify they match the reported activity.
  • Automated Reporting: With one click, the system compiles this multimodal data into a professional PDF or web-based report for donors, complete with "Impact Summaries" and financial breakdowns.

## How we built it

We utilized the latest Google AI Studio tools to build a highly efficient pipeline:

  • Intelligence Layer: Built with Gemini 3 Pro for deep reasoning (matching expenses to project goals) and Gemini 3 Flash for high-speed OCR and image captioning.
  • Audio Processing: We leveraged the Native Audio API to process field interviews and voice logs directly, maintaining the emotional nuance of the beneficiaries.
  • Backend: A secure Firebase architecture to ensure that every "Amanah" record is timestamped, GPS-tagged, and immutable.

## Challenges we ran into

The biggest challenge was Data Variety. Charity bills in rural areas are often handwritten on scraps of paper or formatted irregularly. Initially, standard OCR failed, but by utilizing Gemini 3’s advanced visual reasoning, we were able to train the system to understand "messy" context—like identifying that a handwritten note for "Beras" (Rice) in Indonesia should be categorized under "Food Supplies" in a global report.

## Accomplishments that we're proud of

We successfully achieved Zero-Typing Reporting. A field worker can complete a full distribution report without typing a single word—just by taking photos and speaking to the app. This drastically increases the frequency of reports, giving donors real-time updates rather than waiting months for a manual newsletter.

## What we learned

We learned that transparency is the best currency. When donors see a photo of a receipt matched with a GPS-tagged photo of the delivery, their trust in the NGO triples. We also learned that AI shouldn't replace the human story; it should act as the "secretary" that clears away the paperwork so the human charity workers can focus on helping people.

## What's next for Amanah App: Charity and Funding Report System

Our next milestone is Fraud Detection AI. We are working on a module that can flag duplicate receipts or inconsistent pricing across different regions to ensure that every cent is spent optimally. We also plan to integrate Blockchain ledgers to make every "Amanah Report" a permanent, unchangeable record of good deeds.

Built With

  • gemini
  • geminiapi3
  • googleaistudio
Share this project:

Updates