Decluttered-Dreams

• Project summary
  • The issue we are hoping to solve
  • How our technology solution can help
  • Our idea
• Technology implementation
  • IBM watsonx product(s) used
  • Other IBM technology used
  • Solution architecture
• Presentation materials
  • Solution demo video
  • Project development roadmap
• Additional details
  • How to run the project
  • Live demo
• About this template
  • Contributing
  • Versioning
  • Authors
  • License
  • Acknowledgments

Project summary

The issue we are hoping to solve

Hoarding and severe clutter can severely impact mental health, financial stability and safety in the home. Many people who need help feel shame or don’t know where to start — and existing services (mental health support, cleaning, reselling) are fragmented. People need a single, compassionate, trustworthy channel that coordinates the right mix of professional support and helps individuals reclaim living space and dignity.

How our technology solution can help

Decluttered-Dreams brings together empathetic conversational AI, automated orchestration of multi-provider workflows, and an AI-driven resale/marketplace to:

• Assess the scope of clutter (image + conversational intake),
• Coordinate vetted professionals (organizers, cleaners, psychologists, financial advisors),
• Convert usable items into income via an integrated resale flow,

Our idea

Decluttered Dreams is a holistic, AI-powered platform that manages the lifecycle of clutter:

1. Coordinated Team of Experts (Orchestration): Users get a single point of contact — an empathetic AI agent — that schedules and coordinates multiple providers (psychologists, cleaners, organizers, resellers). The orchestration layer manages bookings, reminders, and communication to reduce friction for vulnerable users.

2. AI-Powered Resale & Circularity: Users upload photos of their space. Our AI (image understanding + classification) identifies sellable items, suggests categories and pricing, and can automatically prepare listings for an integrated community marketplace — returning financial value to the user and keeping usable goods in circulation.

More detail is available in our description document.

Technology implementation

IBM watsonx product(s) used

Featured watsonx products

• watsonx.ai — Used for multimodal analysis (image item recognition, category detection, sellability scoring) and for generating empathetic conversational responses and recommendations. Where used: image processing pipeline, item categorization, pricing suggestion engine, and parts of the conversational agent that require generative understanding.

• watsonx Orchestrate — Used to automate and coordinate multi-step workflows: provider discovery, scheduling, confirmation messages, follow-up checks, and multi-provider task sequencing (e.g., psychologist → organizer → cleaner). Where used: backend orchestration layer that executes appointment workflows and provider handoffs.

Note: If you review our submission code, you will see calls to the watsonx services in server/ (backend connectors) and client-side chat components that integrate with the Assistant.

Other IBM technology used

Additional IBM AI services

• IBM COS Bucket – After an image is uploaded to the Analyze Room function, the images of products identified and returned by Watsonx.ai during analysis are stored in the IBM Cloud Object Storage (COS) bucket. These stored images are then used to list items in the marketplace.

Solution architecture

The project is built with a decoupled frontend and backend, centered around an AI-agentic framework. The architecture separates user-facing concerns from AI/processing services and orchestration.

The platform operates with two distinct user flows:

1. The Consumer Journey

1. Chat with AI Agent: The user begins an intake conversation with the watsonx Assistant-powered chatbot.
2. Needs Assessment: The assistant asks questions, the user selects a Category (Psychologist, Cleaning, Resale, etc.) and may upload images of their space.
3. Define Scope: The user provides Budget and Location (or accepts suggested plans).
4. Get Provider List: Orchestrate queries the provider database and returns ranked matches filtered by location, budget, and trustscore.
5. Review & Select: The user reviews provider profiles (rating, reviews, trustscore).
6. Book Appointment: The user books a provider; Orchestrate coordinates scheduling and reminders.
7. Provide Feedback: After service, the user submits feedback. NLU and the feedback pipeline update provider trustscore.
8. Upload Image: Users can upload an image of a cluttered room, and the WatsonX AI will assist in identifying products within the image that can be resold.
9. Purchase Items: Users can browse and purchase items available in the community marketplace.

2. The Service Provider Journey

1. Login & Authentication: Verified providers access a secure dashboard.
2. Manage Appointments: Providers see requests and can Accept/Decline.
3. Performance Insights: Providers receive AI-generated insights and prompts for reputation improvement.
4. Reputation Management: trustscore reflects service quality, responsiveness, and client feedback.

Presentation materials

Solution demo video

Project development roadmap

Additional details

How to run the project

Below are the basic steps to run the project locally for development and testing.

Prerequisites

• Git
• Python 3.11+ (backend)
• Node.js (v18+) & npm (frontend)
• (Optional) Docker for containerized runs

Clone the repository

git clone https://github.com/pranith-rao/Decluttered-Dreams.git
cd decluttered-dreams

Backend (Server) Setup

# from repository root
cd server

# create & activate virtualenv (example)
python -m venv .venv
source .venv/bin/activate   # macOS / Linux
# .venv\Scripts\activate    # Windows

pip install -r requirements.txt

# create .env template and edit
vi .env
# Add API keys:
# IBM_API_KEY, IBM_PROJECT_ID

# Run the backend
python app.py

• Backend will run at http://127.0.0.1:5000
• API (Swagger/OpenAPI) docs available at http://127.0.0.1:5000/docs/ (if Swagger UI is enabled)

Frontend (Client) Setup

# open a new terminal, from repo root:
cd client

npm install

# create .env template and edit
vi .env
# Add API keys:
# API_URL

# Run the Vite dev server
npm run dev

• The client app will typically be available at http://localhost:5173/

Client helper script (optional)

Some local demo scripts (e.g., local security helper) live in client/scripts/. If used:

cd client/scripts
npm install
# provide required keys into ./wxo_security_config/ per README in scripts
node server.js

Notes

• Several features require valid IBM Cloud service credentials (watsonx services, Watson ML endpoints, TTS/STT). For local testing, you can stub those endpoints or use test keys where allowed.
• See docs/DESCRIPTION.md for sample environment values and demo accounts (if provided).

Live demo

A running demo (if available) and demo credentials are documented in ./docs/DESCRIPTION.md.

See our description document for login credentials and additional demo notes.

---

About this template

Contributing

Please read CONTRIBUTING.md for details on our code of conduct and the process for submitting pull requests.

Versioning

We use SemVer for versioning. For the versions available, see the tags on this repository.

Authors

Pranith Rao

Sanket Phansekar

Purva Naik

Rahul Vishwakarma

• Decluttered-Dreams Team — Project contributors and maintainers. See the repository contributors list for full details.

License

This project is licensed under the Apache 2.0 License - see the LICENSE file for details.

Acknowledgments

• Built for Call for Code / community-driven social impact initiatives.
• Based on common README templates and open-source examples.
• Thank you to the IBM watsonx and Watson developer communities for service documentation and SDKs.