Bridgy: Helping Overwhelmed Communities Choose the Right Tool, Not Another Tool
Source: Dev.to
[](https://dev.to/agastya_khati_f72c89077c8)
*This is a submission for the [DEV Weekend Challenge: Community](https://dev.to/challenges/weekend-2026-02-28).*
---
## The Community
> “Do you know any tool that can actually help with this? I’ve been searching for an hour and now I’m more overwhelmed than when I started.”
A few months ago a friend reached out to me late at night.
He’s a college student—smart, capable, constantly juggling assignments, exams, and expectations. He also has ADHD.
The problem wasn’t a lack of tools; it was the opposite.
- Productivity apps
- ADHD planners
- Focus tools
- Study systems
- Recommendation threads
- Endless “best apps” lists
Each search added another decision, each decision increased the cognitive load. By the time the search ended, nothing had been chosen and the overwhelm had only grown.
The person reading that message understood immediately because he belongs to the same overlapping communities:
- a **college student**
- a **developer**
- someone living with **ADHD**
In these communities the hardest part is often not execution—it’s **starting**.
And starting means choosing. Choosing means evaluating options, which is exhausting when your brain is already overloaded.
What actually helped wasn’t finding *another* tool. What helped was when someone narrowed the field and said:
> “Based on what you’re dealing with *right now*, start here.”
**Bridgy** is built for that moment.
---
## What I Built
**Bridgy** is a transparent, rule‑based resource‑matching web app designed to reduce decision paralysis—not add to it.
### How it works
1. **Input** – Users can either:
- Type a natural‑language description, **or**
- Follow a simple **3‑question guided flow**.
2. **Interpretation** – The app extracts context (community, problem, urgency).
3. **Matching** – It returns **2–3 curated resources**, each accompanied by an explanation of *why* it was selected.
4. **Next Action** – Suggests **one concrete next step** for the user.
### Principles
- No black‑box recommendations.
- No scraping.
- No tracking.
- No stored personal data.
- Every recommendation is **deterministic and explainable**.
AI is used **only** to understand the natural‑language input and to rewrite explanations empathetically; it never decides which resources are shown.
---
## Demo
### Video Demo
*(Insert video embed or link here)*
---
## Code
**GitHub Repository:**
[https://github.com/your‑username/bridgy](https://github.com/your-username/bridgy)
---
*Feel free to explore, give feedback, or contribute!*Bridgy
Transparent, rule‑based resource matching for neurodivergent minds.
Bridgy helps people with ADHD, autism, dyslexia, anxiety, depression, and related conditions find the right tools and support—without black‑box algorithms, data collection, or tracking. Every recommendation is deterministic and explainable.
Table of Contents
- Overview
- Features
- Tech Stack
- Project Structure
- How It Works
- Getting Started
- Environment Variables
- Pages & Routes
- Data Model
- Scoring Engine
- AI Integration
- Design System
- Scripts
Overview
Bridgy is a static‑first, privacy‑first web app. Users describe what they need — either by answering a short guided questionnaire or by typing freely — and Bridgy matches them to a curated set of 32+ resources using a fully transparent, deterministic scoring engine.
- No backend database
- No user accounts
- No analytics
- No data sent to third parties (except an optional LLM call to parse free‑text input)
All matching logic, scoring rules, and data models live in the public repository, so anyone can read and audit the code.
How I Built It
Architecture Philosophy
Bridgy is intentionally:
- Static‑first
- Privacy‑first
- Explainable by design
Consequences:
- No backend database
- No user accounts
- No analytics
- No stored user data
The app works fully even without an API key.
Tech Stack
| Layer | Technology |
|---|---|
| Framework | Next.js 16 (App Router) |
| Language | TypeScript (strict) |
| UI | React 19 |
| Styling | Tailwind CSS v4 |
| Icons | Lucide React |
| LLM Gateway | OpenRouter (GLM‑4.5 → Gemini 2.0 Flash fallback) |
| Deployment | Vercel |
How Bridgy Works
User Input
- Guided mode – select Community → Problem → Urgency
- Free‑text mode – type a natural‑language description
Intent Parsing
- Free text is parsed into a structured intent using GLM‑4.5.
- If GLM‑4.5 is unavailable, it falls back to Gemini 2.0 Flash.
- When no API key is provided, a keyword‑based local parser runs instead.
Deterministic Scoring
| Match Type | Points |
|---|---|
| Community match | +3 |
| Problem match | +3 |
| Urgency match | +1 |
| Human support (when requested) | +2 |
The same input always produces the same ranked output.
Transparent Results
- 2–3 matched resources are shown.
- Each result includes a “Why this?” explanation.
- A single, clear next action is highlighted.
Why These Decisions Matter
Bridgy deliberately avoids:
- Web scraping
- Live search
- User tracking
- Black‑box AI recommendations
For people who are already overwhelmed, trust and clarity matter more than clever automation. Every design decision aims to reduce cognitive load, not increase it.
Closing
Technology has done a great job building tools for communities, but it has done a poorer job helping people navigate between them.
Bridgy is a small attempt to bridge that gap—not by adding more options, but by helping someone take one clear step forward when everything feels like too much.
Sometimes, that’s enough to change the direction of a day.