DataDon

The need

Teams often depend on a technical person for recurring data questions, while individual AI conversations create inconsistent definitions of revenue, margin or performance. Wider access introduces another problem: deciding which sources and records each team should be able to query.

What Tech Local built

We built DataDon as a shared analytics workspace for connected databases and spreadsheets. A technical owner chooses datasource scope, documents business context and saves reusable queries with parameter defaults. Team members ask questions in plain language, while the application checks workspace access, enforces read-only queries and validates allowed table scope. The built-in analyst can plan a sequence of queries, inspect results and continue an investigation. Results, charts, exports and dashboards make the answers usable beyond the conversation.

Important workflows

Establish the shared context

Connect PostgreSQL, SQL Server or a spreadsheet, choose the relevant data scope and describe the terms, entities and business rules the team uses.

Answer repeatable questions

Save a known query with defaults, run it for a new date range or ask a follow-up. Inspect the returned results and export them for further work.

Investigate and retain useful results

Let the built-in analyst plan multiple read-only queries and inspect the results. Add useful query-backed charts to dashboards; charts saved directly from chat remain snapshots unless linked to a query.

Implementation

• Workspace roles and access controls
• Selected schema and table scope
• Reusable SQL query templates
• Shared business context
• PostgreSQL, SQL Server and spreadsheets
• Built-in multi-step analyst
• Charts, exports and dashboards

Build stack

Next.js · React · TypeScript

Application, workspace interface and server routes

Tailwind CSS · Radix UI

Responsive application components

Supabase · PostgreSQL

Authentication, workspace metadata and access policies

Python · FastAPI

Datasource connector and read-only SQL execution

OpenAI Responses API

Natural-language analysis and multi-step query planning

Recharts · Vercel · Fly.io

Charts, web hosting and independent workers

Business knowledge the whole team can reuse.

Shared context makes conversational analytics useful across a team.

Shared definitions

Keep business definitions and trusted queries together, beyond individual prompts.

Controlled access

Control access through workspace roles and selected schemas and tables.

Repeatable questions

Run the same query with new parameters for a consistent starting point.