ByteScan

Agentic Data Intelligence Pipeline

Powered by Google Gemini 2.5 Flash & OpenAI GPT-4o-Mini Failover

---

What is ByteScan?

ByteScan is a beautiful, highly-optimized, multi-document analysis processing platform. It ingests massive swaths of data (documents, folders, code, JSON), passes them through a sophisticated multi-stage AI extraction loop, and yields highly-actionable intelligence directly into an interactive dashboard.

Multi-File Architecture 🚀

ByteScan allows you to cross-reference multiple PDFs, CSVs, or text blobs dynamically constraint by dynamically distributing token space across nodes so it never hits Quota failures.

Features

• Multi-File & Folder Ingestion: Securely process an entire folder block at once using browser webkitdirectory mounts.
• Fail-Safe Processing: If Gemini reaches token-exhaustion (HTTP 429), ByteScan silently and instantaneously shifts routing arrays to OpenAI GPT-4o-Mini via an embedded polyfill adapter to gracefully complete your intelligence extraction! No 429 timeouts!
• Sleek Mobile UI: Enforced iOS safari zoom-defenses and layout constraints to guarantee perfect mobile readability.

Local Development

1. Clone the repository:

   git clone https://github.com/mew228/ByteScan.git
   cd ByteScan

2. Install dependencies:

   npm install

3. Configure the environment: Create a .env file in the root matching .env.example:

   VITE_GEMINI_API_KEY="your_free_tier_gemini_key"
   VITE_OPENAI_API_KEY="your_openai_failsafe_tier_key"

4. Launch the platform:

   npm run dev

Production

This project has Vite-specific build configurations enabled to support large modular deployment layers to Vercel correctly. Compile safely:

npm run build

Technologies Built Upon

• React 18
• TypeScript
• Vite
• Framer Motion
• @google/genai & OpenAI Node SDK

Data pipelines built for scale.