TensorLake Expense Explorer 🍎🚀

Expense Explorer is a premium, cloud-native financial intelligence suite that transforms messy PDF statements into structured, actionable insights. Powered by TensorLake Cloud and Gemini-3-Flash.

✨ Key Features

• Apple-Style Dashboard: A minimalist, high-daylight UI for seamless interaction.
• Deep Financial Extraction: Beyond dates and amounts—identifies merchants, subscriptions, payment methods, and enriched tags.
• Gemini-3-Flash: High-performance extraction and conversational intelligence.
• Deduplication: Intelligent iterative ingestion skips already-processed transactions.
• SQL Reliability: Integrated tenacity retry logic for resilient database operations.
• Cloud-Native Proxy: A secure server-side proxy handles authentication and bypasses browser security constraints.

🏗️ Architecture

• AI Models: Gemini-3-Flash (via TensorLake Applications).
• Backend API: TensorLake V2 REST API + Application Framework.
• Database: Neon Postgres (Serverless) for robust financial records.
• Web UI: Vanilla HTML/JS with a refined Apple-inspired aesthetic.

🛠️ Quick Start

1. Requirements

Ensure you have the following in your .env file:

TENSORLAKE_API_KEY=YOUR_KEY
GEMINI_API_KEY=YOUR_KEY
DATABASE_URL=YOUR_NEON_POSTGRES_URL

2. Configuration & Deployment

Push secrets to TensorLake and deploy your cloud applications:

tensorlake secrets set DATABASE_URL=...
tensorlake secrets set GEMINI_API_KEY=...
tensorlake deploy workflow.py

3. Ingest Statements

Dashboard (Individual)

Launch the dashboard and drag/drop your files.

CLI (Batch)

Use the batch tool for bulk processing:

python verify_flow.py

Note: Ensure your PDF statements are in ~/Downloads/Credit_Card_Statements.

4. Launch the Dashboard

Use the built-in launch script:

chmod +x launch_dashboard.sh
./launch_dashboard.sh

Open http://localhost:8000 in your browser.

Built with ❤️ using TensorLake.