Skip to content

Charity Planner

Most people want to be more generous but don't know where to begin—there are nearly two million charities in the US alone.

Core trick

Conversational interview captures stated values, then an agent researches charities and assigns dollar allocations with a cited source for each.

Tags

Description

Most people want to be more generous but don't know where to begin—there are nearly two million charities in the US alone. Charity Planner turns a quick interview into a researched, transparent giving actionable strategy: the kind wealthy families pay advisors thousands for, but this time free for everyone—thanks to AI.
acrosa/charity-planner★ 00 forksTypeScriptCreated Jun 13, 2026 · last push Jun 13, 2026

README excerpt

# Charity Planner

_Turn your values into a clear plan for giving._

Most people want to be more generous — they just don't know where to start. With
2M+ U.S. nonprofits, choosing where to give is overwhelming. Wealthy families
solve this by hiring philanthropic advisors who charge thousands a year to build
a giving strategy and a charitable portfolio.

**Charity Planner makes that service free for everyone.** A short, conversational
interview turns your explicit preferences into a researched, transparent,
actionable giving strategy — philosophy → strategy → a charity portfolio with
exact dollar allocations → a plan → an educational Daffy portfolio comparison —
with the reasoning and source behind every charity and every figure, and a clear
handoff to act on the plan through [Daffy](https://www.daffy.org).

> Built for the **Claude Build Day** hackathon.
> **Live:** https://charity-planner.fly.dev · try the [quick interview](https://charity-planner.fly.dev/plan?mode=quick).

## How it works

1. **Interview** — a short structured conversation (quick: 3 questions, or full)
   captures causes, goals, timeline, constraints, and risk preferences. Facets are
   only ever captured from what you actually say — nothing is fabricated.
2. **Retrieve** — your philosophy is expanded into retrieval queries and exclusion
   predicates; candidates are recalled from a 10k-charity corpus via vector search
   (`pgvector`) ∪ a trigram name leg, then reranked (Voyage `rerank-2.5`),
   diversified

Team