Extractly

Extractly is a Streamlit app for defining document schemas, classifying incoming files, and extracting structured metadata with traceable runs. It is designed to feel like a client-ready “Document Metadata Extraction Studio” with clear workflow steps and demo-friendly outputs.

Quickstart

1. Create a .env file with your API keys:

OPENAI_API_KEY=your_key_here

2. Install dependencies (for example with uv or pip).
3. Run the app:

streamlit run Home.py

App Navigation

• Home: Landing page with product highlights and quick CTAs.
• Schema Studio: Create, edit, validate, and export schemas with a live JSON preview.
• Extract: Upload single or batch documents, select routing, and run extraction.
• Results: Browse run history, inspect JSON/table outputs, export JSON/CSV.
• Settings: Review models, retries, and environment configuration.

Schemas

• Prebuilt schemas: schemas/prebuilt_schemas.json
• Custom schemas: schemas/custom_schemas.json

Custom schemas override prebuilt ones when names overlap.

Runs

All extraction runs are stored in data/runs/ with input filenames, output JSON, and logs.

Demo Script

1. Open Schema Studio and review a prebuilt schema (e.g., Invoice Demo).
2. Go to Extract, choose routing, and upload a sample document from data/sample_docs/ or data/imgs/.
3. Run extraction and open Results to review output and export JSON/CSV.

Notes

• Update models and retries via .env using the EXTRACTLY_* variables.
• Keep API keys in environment variables only; do not hardcode secrets.