Databricks Community

Viral0216 · a week ago

Hi everyone,

I built Intelligence Studio, an open-source workbench that lets you browse, test, analyse, and integrate with 640+ Databricks REST APIs -- all from one interface. No more juggling docs, curl commands, Postman collections, and multiple browser tabs. I'd love your feedback.

The Problem

Working with Databricks APIs today means:

640+ endpoints scattered across documentation -- hard to find the right one
Manual testing with curl, Postman, or custom scripts for every request
Token management across multiple workspaces is tedious and error-prone
Cryptic error messages with no guidance on root cause or fix
No built-in SQL editor for quick queries without switching to another tool
Writing boilerplate code every time you want to call an API from your app

What Intelligence Studio Does

A single app that combines API explorer, SQL editor, AI assistant, and code generator -- all connected to your Databricks workspace.

Core Features

Feature What It Does

API Catalog	640+ Databricks endpoints (500+ workspace + account APIs), searchable and organised
Request Playground	Auto-generated request bodies, method/path builder, response formatting, pagination
SQL Query Editor	Full editor with syntax highlighting, Unity Catalog browser (Catalog > Schema > Table > Column)
Azure Multi-Workspace Login	OAuth-based progressive selection (Tenant > Subscription > Workspace)
Request History	Full replay capability with favourites and version tracking
Data Visualisation	12 chart types with auto-detection from API responses
Code Export	Python, cURL, JavaScript, TypeScript, Go, PowerShell -- one click
Integration Export	Postman, Insomnia, OpenAPI 3.0, GitHub Actions

13 AI Assistant Features

This is where it gets interesting. Intelligence Studio has 13 AI-powered features built throughout (not bolted on), powered by Databricks Foundation Models:

Data Q&A -- Ask questions in plain English, get SQL generated and executed ("Show me all tables in catalog X with row counts")
Find Endpoint -- Describe what you want in natural language ("list all running jobs") and get the right API endpoint
Analyse Response -- Pattern detection and insights from any API response
Error Analysis -- Root cause diagnosis with step-by-step fix suggestions for cryptic API errors
Code Generation -- Export tested requests as production-ready code in 6 languages
Workflow Builder -- Chain multi-step API calls where one step's output feeds into the next
Agent Chat -- AI agent that calls Databricks APIs on your behalf
Test Data Generator -- Schema-aware synthetic test payloads
API Docs -- Browsable, searchable endpoint documentation
Query Builder -- Natural language to SQL conversion with execution
Data Visualisation -- 12 chart types with auto-detection, dashboards, lineage graphs
AI Scripting -- Python script generation with sandboxed execution
Prompt Manager -- Customise 14 system prompts for your organisation

Tech Stack

Frontend:   React 18 + TypeScript 5 + Vite 6 + Zustand + Tailwind CSS 4 + Recharts
Backend:    FastAPI + Uvicorn + httpx + Pydantic v2 + azure-identity
Desktop:    Electron 28 (macOS arm64 + Windows)
CLI:        Python Click + Rich
AI:         Databricks Foundation Models (Llama, Claude, Gemma)

Security First

No server-side token storage -- each request carries its own credentials, tokens are never persisted
Sandboxed script execution -- AI-generated Python scripts validated against blocklist before running
Pydantic v2 validation -- all inputs validated before processing
Per-request auth -- backend acts as an authenticated proxy, nothing stored

Run It Your Way

Mode How

Web App	make dev -- React + FastAPI on localhost
macOS Desktop	Download .zip, double-click to launch
Windows Desktop	NSIS installer or portable .exe
CLI	make cli-install for terminal workflows

Quick Start (5 minutes)

git clone https://github.com/viral0216/Intelligence-Studio.git
cd Intelligence-Studio
make install
make dev
# Open http://localhost:5173
# Configure Databricks host + token in Settings

Requirements: Python 3.11+, Node.js 18+, Databricks workspace + PAT

Enterprise Features

14 feature flags for team policies and governance
Cost tracking per AI model call with token usage metadata
14 Customisable AI prompts -- tailor the AI to your organisation's context
Full export suite -- PDF, Word, Markdown, Excel, CSV, JSON

By the Numbers

Metric Value

API endpoints cataloged	640+
AI assistant features	13
Code generation languages	6
Export formats	10+
Chart types	12
Feature flags	14
Customisable prompts	14
Platforms	Web, macOS, Windows, CLI

What I'm Looking For

Feedback -- Does this solve real pain points in your Databricks workflow?
Feature requests -- What's missing? New chart types? More export formats?
Bug reports -- It's v1.0, so expect some rough edges
Contributors -- PRs welcome! Areas: new endpoint presets, AI prompt templates, chart types, tests, docs