ReachOut is a project that facilitates micro-lending to users in the most remote areas of the world. Lenders can make numerous small, high-return investments that allow borrowers to purchase crucial items such as fertilizer, young animals, agricultural tools, or spare parts.


Our Android app provides an extremely accessible way to create proposals for loans, and to receive funds.

  1. The user interface requires zero literacy.
  2. An entire community can share a single device.
  3. Proposals created offline can be synched with our database when users travel to a city with internet.

Community Credit

Our lending model allows communities to build credit from scratch.

  1. Community leaders can endorse requests for loans.
  2. Endorsed loans that are repaid successfully increase the leader’s ReachOut credit score.
  3. This allows subsequent investors to invest in the community with confidence, without requiring the local nonprofits to underwrite loans

How it works

  1. Text-free Login: Users create accounts by taking pictures of themselves, and login by clicking on their picture and entering a pattern lock.
  2. Text-free Proposal Generation: Users are given a sequence of audio prompts about their requested loan and how they will use it. For each prompt, user responses are recorded as text using Google’s speech-to-text API. Users can play back their responses to verify that transcripts are accurate, and attach pictures to support their proposals.
  3. Database: Proposals are tracked in DynamoDB on Amazon Web Services, and stored data is used to calculate credit scores.
  4. Easy access to funds: Accounts and transfers are managed automatically using Capital One’s Nessie API. To withdraw cash from their loan, users allow a banker to type an account number into their their withdrawal screen, and the full amount of the loan is transferred.

Technologies used

Back-end: Capital One’s Nessie, DynamoDB, Amazon Web Services

Front-End: Android, Google speech-to-text and text-to-speech, Pattern lock, Camera integration

Challenges we ran into:

  1. Lots of moving parts, lots of tools to learn.
  2. Storing and retrieving data in DynamoDB required us to flatten our model, but we were able to do it without losing information.
  3. Creating a web client that interfaced with DynamoDB and Nessie was very difficult.
  4. Tracking a proposal through different states created a lot of unexpected difficulties.

Next steps

  1. Google Maps Integration: Automatically display satellite imagery of the borrower’s location.
  2. Localization of accounts, allowing leaders to view proposals within a certain radius.
  3. Create alternate prompts for different languages.
  4. Create more meaningful icons.
Share this project: