This is a submission for Weekend Challenge:Generosity Edition
What I Built
During crises and natural disasters, traditional relief faces some issues, such as aid arriving with the wrong resources or a lack of transparency while handling the funds. Donors rarely know whether the funds actually reach on the ground or not. Here is the comes KindLink as a solution with a decentralised autonomous disaster response system to designed to coordinate with multi-modal generosity in real time. It unifies monetary and non-monetary aid(food, ration, transport) with on-chain transactions. It also provides an impact receipt to the donor so they will actually know where the funds are utilized.
Demo
In English :-https://youtu.be/wgNX2KGTtFU?si=Y8vL6usT6jUGNDPK
In Hindi :- https://youtu.be/tZvU03fhlrw?si=eHUalW5M0x3-4DjO
Code
Github repo :- https://github.com/GamersStop/KINDLINK-The-Autonomous-Generosity-Grid
/backend :- Used Python FastAPI ,SQLModel and python integration with solders and snowflakes
/Frontend :- Used next.js for the responsive screen across all the devices.
docker-composer.yml :- It uses for the smooth deployment
How I Built It
As explained in the video, we have 3 flows in the application to achieve that, we used the below technology
Backend on FastAPI & PostgreSQL:-
The backend is built with Python, FastAPI, and SQLModel, running inside a Docker container for seamless deployment. It handles all incoming requests from the frontend, while dedicated API endpoints process tasks like SOS and plea analysis. To organize and route requests to their specific endpoint files, we utilized FastAPI routers. Additionally, all database schemas and data structures are defined in the models directory. If volunteer click on the pay by sol button then those sponcership is directly implement via Solana devnet using the python native solders. It will generate the Solana transaction signature, which you can directly verify with the Solana Explorer. For admin dashbord we did integration with snowflakes to make charts from analytical database which provides us the great trend analysis so admin will know hows the trends were working hows the funds flows vs hows the commitment of efforts were flows with their solana on chain transactions.
Frontend Build with Next.js
We chose Next.js for the frontend with the main motive of providing a responsive ui across all devices.All the admin related UI is in /admin folder and for session maintenance, we use the components to guard on our two ngo and giver flow and use the HTTP only sessions.
Prize Categories
Best Use of Google AI
- Unstructured Triage Extraction: Used the Gemini API to parse natural language distress calls and unstructured field pleas into standardized, categorized relief SKUs (rations, transport, medical supplies, physical labor).
- Automated Urgency Scoring: Gemini evaluates contextual risk signals to assign priority ratings, enabling instant routing on the triage grid.
Best Use of ElevenLabs
- Field Audio Dispatch: Converted critical mission updates and urgent resource alerts into natural, lifelike voice broadcasts using ElevenLabs.
- Hands-Free Ground Coordination: Enabled field volunteers and first responders to listen to real-time status shifts and dispatch instructions without needing screen access.
Best Use of Snowflake
-
Transactional-to-Analytical Pipeline: Synchronized operational state from PostgreSQL into the
KINDLINK_ANALYTICSdatabase viasnowflake-connector-python. -
Multi-Modal OLAP Analytics: Built analytical SQL views (
V_MODAL_AID_DISTRIBUTIONandV_MISSION_CAPITAL_VELOCITY) to track resource allocation and capital velocity ratios across disaster zones without running into trial function restrictions.
Best Use of Solana
-
Autonomous Escrow Rails: Implemented trustless Devnet funding pipelines directly in the backend using
soldersand Solana RPC. - Verifiable Receipts: Every capital commitment broadcasts an immutable on-chain transaction that donors can independently audit via Solana Devnet Explorer.

Top comments (0)