AInalise

Connect data sources to AI analysis.

Identified technologies

Python / FastAPIReactFlutter

The problem

Working with data from different sources needs more than a query screen. Each database has its own query rules, access controls and connection method.

The solution

AInalise connects data sources to AI analysis. The API, connectors and local agent handle access. Web and mobile apps present the analysis.

What makes up the solution

Python/FastAPI API with contracts, access controls, metadata and adapters for external sources.
View technical details →
Python agent for local network sources, with metadata collection and read-only query checks.
View technical details →
React/TypeScript system and Flutter app use the API without copying the connectors to the device.
View technical details →

Project scope The platform’s PostgreSQL database is separate from the queried sources. Central or agent access depends on the allowed connection mode and each environment.

My contribution

I designed AInalise and developed its architecture, API, connectors, agent, web system and mobile app. The work separates data-source access from the analysis experience.

Architecture overview

A Python API, connectors and an agent connect data sources to web and mobile analysis.

Technical structure

Technology by module.

Web, API, app and local agent have different roles in analysis and authorized data access.

What each part does and which technologies support it. A module in the code does not mean all its features are complete or published.

01 / Existing code

System

ReactTypeScriptViteReact Query

Analysis and administration interface.

Routes, query state and components present data and API-authorized operations, without direct browser access to client databases.

02 / Existing code

API and connectors

PythonFastAPISQLAlchemyPostgreSQL

Central API and connectors.

Feature-based services organize contracts, permissions, metadata and queries. Adapters handle external data-source dialects and connection modes.

03 / Existing code

Mobile app

FlutterDartRiverpodDio

Mobile platform interface.

State and HTTP communication have separate layers. The app uses API contracts without copying server database connectors.

04 / Existing code

Agent

PythonTyperFastAPISQLAlchemyhttpx

Access to local network data sources.

Agent with configuration, metadata collection, read-only query checks and communication with the central API.

How the parts connect

Web and app use the API. Sources can be reached through a central connector or local agent, based on authorized connection mode and environment capabilities.

Data and integrations

PostgreSQL stores platform data. PostgreSQL, MySQL, SQL Server and Oracle are external-source types. Azure SQL/Fabric have their own authentication modes.

Quality and operations

Separate API, web and agent checks, documented AWS hosting and dedicated workflows. Agent installation, platform pairing and connector operation are separate checks.