Apppublic

sbd00/Echo-Archive

sourceHugging Facegpl-3.0updated 4mo agoView on Hugging Face

0likes

App README

Echo Archive

Echo Archive is an AI-powered reflective archive designed to help users connect with artwork they may have never explored before.

For many people, artwork can feel historically distant, emotionally inaccessible, or difficult to interpret without formal artistic knowledge. Traditional digital museum experiences often prioritize metadata, search, and categorization, but rarely help users build a personal connection to the work itself.

Echo Archive addresses this problem by combining archival retrieval with generative AI interpretation.

System Architecture

text id="e04x6v"

User Reflection Input
        ↓
[Art Institute of Chicago API]
        ↓
Artwork Retrieval + Metadata Filtering
        ↓
Streamlit Artwork Display
        ↓
[Gemini 2.5 Flash]
Curatorial Description
        ↓
[Gemini 2.5 Flash]
Reflective Interpretation
        ↓
User or Gemini-Assisted Reflection
        ↓
Email Archival System

Key Design Decisions

Retrieval Before Generation

Artwork retrieval is grounded in real museum records and filtered for richer historical metadata.

Layered Interpretation

The system separates:

—artwork retrieval
—curatorial description
—interpretation
—personal reflection

to create a paced reflective experience.

Human + AI Co-Authorship

Users can either write their own reflections or use Gemini-assisted drafting.

Tech Stack

—Streamlit
—Python
—Google Gemini 2.5 Flash
—Art Institute of Chicago API
—Resend Email API
—dotenv

Core Features

—Randomized artwork retrieval with historical metadata filtering
—AI-generated curatorial descriptions
—Reflective interpretation pipeline
—User-authored or Gemini-assisted reflections
—Email archiving of encounters and reflections

Run Locally

bash id="v9m65e"

pip install -r requirements.txt
streamlit run app.py

Create a .env file with:

env id="f3hkgm"

GEMINI_API_KEY=your_key_here
RESEND_API_KEY=your_key_here