GradientHarbor

Appearance

Platform Overview ​

GradientHarbor is a BI and data analysis platform that combines AI-powered analytics with traditional dashboard building. It lets you connect your databases, ask questions in natural language, and create shareable dashboards — all from a single interface.

Key Capabilities ​

AI Data Agent ​

The conversational AI agent is the core of GradientHarbor. You ask questions in plain English, and the agent:

  1. Searches your data catalog to find relevant tables
  2. Writes and executes SQL queries against your connected databases
  3. Returns results as tables, charts, or narrative answers
  4. Can create dashboard cells directly from conversation results

The agent understands your schema, column descriptions, and table relationships — so you don't need to know the exact table names or SQL syntax.

Dashboard Builder ​

Build multi-page dashboards with a variety of cell types:

  • Tables — Query results displayed in sortable, paginated tables
  • Charts — Bar, line, area, pie, radar, scatter, and treemap visualizations
  • KPIs — Key metrics displayed as large numbers with optional comparisons
  • Filters — Interactive inputs that parameterize queries across cells
  • Text — Rich text content for headers, descriptions, and annotations

Dashboards can be built manually or with AI assistance — describe what you want, and the agent creates the cells for you.

Data Catalog & ERD ​

When you connect a database, GradientHarbor automatically introspects your schema and builds a searchable catalog of tables, columns, indexes, and foreign keys. The Entity Relationship Diagram (ERD) visualizes table relationships.

Search across your dashboards, tables, and saved queries using natural language. Results are ranked by relevance and usage frequency.

Architecture at a Glance ​

┌─────────────────────────────────────────────┐
│              Browser (React SPA)            │
│  Chat ─ Dashboards ─ Catalog ─ Search      │
└──────────────────┬──────────────────────────┘
                   │ GraphQL
┌──────────────────▼──────────────────────────┐
│            Backend (FastAPI + Strawberry)    │
│  Agent ─ Query Execution ─ Catalog Sync     │
└──────┬───────────┬──────────────────────────┘
       │           │
  ┌────▼────┐ ┌────▼────┐
  │ Your DB │ │   S3    │
  │ (PG/BQ) │ │ (cache) │
  └─────────┘ └─────────┘
  • Frontend: React + TypeScript, deployed on Cloudflare Workers
  • Backend: FastAPI with Strawberry GraphQL, deployed on AWS ECS
  • Communication: GraphQL is the primary API (REST only for OAuth callbacks)
  • Worker: Async job consumer (RabbitMQ) handles catalog introspection and background tasks

Next Steps ​