​
Brighthive Platform Overview

​
High level architecture

​
Platform Architecture Overview

Our platform is designed for data-driven organizations looking to integrate disparate data sources, manage ingestion pipelines, and transform and explore data seamlessly through a modern, conversational interface. Below are our key differentiators and technologies we employ in our stack.

​
💡 Key Differentiators

  • Unified Experience: Configure ingestion, transformation, and warehouses all from a single UI.
  • Conversational Querying: Explore data through natural language with BrightBot.
  • Modular Integration: Seamless compatibility with your existing cloud infrastructure.
  • Graph-powered Governance: Track and manage everything through a powerful knowledge graph.

​
🧰 Technologies

A complete list of technologies used across the platform, grouped by category.

​
🖥 Frontend

  • React – Component-based UI framework.
  • GraphQL Apollo Client – Declarative data fetching and state management.
  • TypeScript – Static typing for better development experience.

​
🧠 Conversational Interface

  • BrightBot – Internal chatbot service.

​
☁️ Backend

  • GraphQL – Unified API layer for frontend and backend communication.

​
📂 File Upload & Storage

  • Amazon S3 – Object storage for CSV, PDF, image, video uploads.
  • Neo4j – Metadata store and indexing for uploaded assets (Data Catalog).

​
🔌 Ingestion & Integration

  • Airbyte (EC2-hosted) – Open-source data integration platform.
  • Custom Ingestion Service – API layer that creates and manages Airbyte sources/connections.

​
🧊 Data Warehousing

  • Amazon Redshift – Columnar data warehouse.
  • Snowflake – Cloud-native data warehouse.

​
🔄 Transformation

  • DBT Cloud – Transformation-as-code platform for analytics engineering.

​
🔍 Metadata & Data Graph

  • Neo4j (EC2-hosted) – Central knowledge graph representing:
    • Files
    • Sources
    • Pipelines
    • Warehouses
    • Transformations
    • Other application components.
  • Cypher Query Language + GraphQL OGM - for querying neo4j.