OmniSight: Autonomous Corporate Strategy & Disruption Radar

Multi-source competitive intelligence engine synthesizing live web telemetry, audio transcripts, and knowledge graphs into dynamic strategy dashboards.

---

Overview

Market strategy and competitive analysis workflows often require manual synthesis across earnings call audio, pricing changes, hiring velocity, and web telemetry.

OmniSight automates this analysis through a concurrent agent pipeline:

1. Live Web & Search Scraping: Extracts recent pricing pivots, hiring trends, and customer sentiment signals via SERP APIs and headless web scrapers.
2. Audio Intelligence: Transcribes executive earnings calls or product presentations with speaker diarization.
3. Graph Memory Synthesis: Persists entity relations and historical market shifts into a local vector/knowledge graph via Cognee.
4. Interactive Visualization: Formats findings into dual-layered radar comparisons and structured intelligence battlecards.

---

Architecture Flow

sequenceDiagram
    autonumber
    participant Client as Analyst Dashboard
    participant API as FastAPI Backend (Asyncio Engine)
    participant Scraper as Bright Data SERP / Web Scraper
    participant Audio as Speechmatics Diarization API
    participant Graph as Cognee Graph Memory
    participant LLM as Llama-3 / Gemini Reasoning

    Client->>API: Initiate Company Audit (Entity + Audio Source)
    par Concurrent Ingestion
        API->>Scraper: Query Search Indices & Scrape Landing Pages
        API->>Audio: Submit Audio Stream for Batch Transcription
    end
    Scraper-->>API: Extracted Web Metrics (Pricing, Job Postings, Reviews)
    Audio-->>API: Diarized Speaker Transcripts
    API->>LLM: Synthesize Multimodal Context into Structured JSON
    LLM-->>API: Normalized Competitive Vector (Scores & Risk Factors)
    API->>Graph: Ingest Entities and Update Historical Market Graph
    API-->>Client: Return Radar Dataset & Strategic Battlecard

Loading

---

Core Capabilities

• Concurrent Agent Ingestion: Parallel asynchronous coroutines execute web scraping and batch audio processing concurrently to reduce pipeline latency.
• Multimodal Diarization: Extracts strategic signals, executive quotes, and sentiment directly from multi-speaker audio recordings.
• Persistent Market Graphs: Maintains cross-scan entity relationships in a local vector/knowledge graph.
• Dynamic Radar Analytics: Produces standardized competitive vectors across hiring velocity, product innovation, pricing pressure, and customer sentiment.

---

Repository Structure

.
├── static/                   # Frontend assets, Tailwind CSS, Chart.js radar views
├── templates/                # HTML dashboard layouts
├── server.py                 # Core FastAPI server and asynchronous pipeline
├── requirements.txt          # Python dependencies
└── README.md

---

Getting Started

Prerequisites

• Python 3.10 or higher
• API credentials for Google Gemini (or Groq), Bright Data, and Speechmatics

Local Setup

1. Clone the repository:

   git clone https://github.com/HamzaKhanBUIC/OmniSight-Autonomous-Corporate-Strategy-Disruption-Radar.git
   cd OmniSight-Autonomous-Corporate-Strategy-Disruption-Radar

2. Install dependencies:

   pip install -r requirements.txt

3. Configure environment variables: Create a .env file in the root directory:

   BRIGHT_DATA_API_KEY=your_bright_data_key
   SPEECHMATICS_API_KEY=your_speechmatics_key
   GEMINI_API_KEY=your_gemini_key
   GROQ_API_KEY=your_groq_key

4. Launch the backend server:

   python server.py

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

---

Deployment Guidelines

When running on Linux-based container runtimes (such as Google Cloud Run or AWS ECS), ensure the container entrypoint invokes python server.py directly and allocate at least 1 GB of memory for local vector indexing during audio processing.

---

License

This project is licensed under the MIT License. See LICENSE for details.