PIL ‘26 Spotlight: Tech Bono for affordable housing– Building a custom CRM for a NYC Community Land Trust Nonprofit

How do you build a custom CRM for a Community Land Trust that is affordable, flexible, and maintainable?

This summer, our team participated in hackNY's Public Interest Lab (PIL), which brought together technologists to tackle technology challenges facing nonprofit and civic organizations. Through PIL's tech-bono model, we partnered with the East Harlem/El Barrio Community Land Trust (EHEBCLT), an organization working to preserve affordable, community-controlled housing for low-income households.

EHEBCLT manages multiple residential properties, but this information is scattered across spreadsheets and paper records. Our goal was straightforward: build a CRM that would give staff one centralized place to manage this information and make their day-to-day work easier.

Before we could build anything, we had to understand who we were actually building for. One of our first challenges was understanding how a Community Land Trust actually operates. It is a pretty unique housing model and one that most of us had not encountered. Getting the data model right meant first understanding the language, learning concepts like rent stabilization, income certification thresholds, and mutual housing, and understanding how they connect to each other in practice. A single lease in the community land trust model can be subsidized through a program like SCRIE, DRIE, or Section 8. Section 8 itself further breaks down into Project Based Voucher (PBV) and Housing Choice Voucher (HCV) types, and a unit can have more than one subsidy program at once. Beyond traditional lease data like unit type and monthly rent, EHEBCLT needs to track information on each subsidy attached to a unit, along with the income certifications tied to it. The complicated nature of the data relationships in the CLT model meant that we had to first work to understand the problem at hand and ensure the database schema was accurate before building the software itself. 

Rather than building from scratch, we started by researching existing open-source CRMs. Public interest technology isn’t about engineering the most impressive tech stack; it’s about meeting organizations where they are. For a team like EHEBCLT with no dedicated technical staff, an intuitive, maintainable off-the-shelf solution was far more valuable than a complex custom build.

We narrowed our search to three primary platforms, but each had key limitations:

• CiviCRM: Open-source and mission-aligned, but its rigid data models, dated interface, and poor mobile responsiveness demanded more custom development than our ten-week fellowship allowed.

• Yardi: Powerful for property management, but per-unit pricing and monthly minimums made it an expensive, overly complex choice. It also does not map cleanly to the CLT model.

• Twenty: Highly flexible with low-cost self-hosting, but its enterprise-first architecture didn't map cleanly to EHEBCLT's specific needs. Specifically, it does not support authentication with scoped permissions, which prevents a resident portal from being built with it in the future. 

Ultimately, we chose to build a lightweight, custom platform tailored directly to EHEBCLT’s workflows. We selected PocketBase for the backend, which allowed us to design flexible schemas and APIs while keeping infrastructure overhead and hosting costs near zero. Paired with a React frontend, we automated deployments through GitHub Actions so future maintainers can update the system friction-free. The MVP covers EHEBCLT's core data, including buildings, units, tenants, and household members, providing an intuitive foundation with plenty of room to grow.

Throughout this fellowship, AI coding tools helped us prototype rapidly, explore different architectures, and iterate as needed. However, that increased speed only reinforced that velocity means nothing if you build the wrong thing. While AI lowers the cost of generating code, it does not change developer accountability or replace the work of understanding users' lived experiences. When building for a community organization, our ultimate responsibility is to deliver a stable, intuitive tool built for the realities of the people who rely on it.

For all of us, hackNY’s PIL was a rewarding and meaningful experience. Beyond just building a CRM, it was a chance to develop technology that addresses real challenges for civic organizations and everyday New Yorkers. Balancing the fellowship alongside full-time work pushed us to communicate clearly, scope carefully, and show up consistently. Our biggest lesson was that public interest tech is rarely a one-size-fits-all problem. EHEBCLT's unique data structure meant adapting to affordable housing workflows that a non-technical staff could easily manage, while exploring creative approaches like leveraging MCP servers to reduce developer effort. This summer highlighted for us why public interest technology matters: when tools significantly improve organizational operations, they make public resources more accessible and better serve the communities that depend on them.

Check out our repo here: https://github.com/shreeyachand/ehebclt-crm/ 

Team members: Ayesha Ahmad h’26, Shreeya Chand h’26, and Hayun Jung h’26

Mentor: Ray Berger h’18

Subject Matter Expert: Maria Paula Moreno Vivas