qData: an open-source data middle platform for enterprise governance

A data middle platform, not a data tool

qData is described as an open-source data middle platform for enterprise data governance and data development scenarios. The README is explicit about the scope: ETL data integration, data development, data modeling, metadata management, data quality, data assets, API data services, and AI-powered data Q&A, with support for common databases including MySQL, DM8, Oracle, SQL Server, Kingbase8 and Doris.

The intended outcome is described in business terms. qData helps enterprises complete data access, cleansing and transformation, asset cataloging, quality checks, API publishing and Text2SQL analysis, and it can serve as a foundation for enterprise data middle platforms, data governance platforms, ETL platforms and data service platforms, and it is also described as suitable for secondary development and feature extensions.

So the pitch is a platform, and the practical implication is that this is a whole application rather than a component. The repository is a Maven project with a server module, a UI module, a dozen business modules and a docker directory, and the stack badges list JDK 1.8+, Spring Boot 2.5.15 and Vue 3.4.31.

The project is published in three README languages, English, Simplified Chinese and Japanese, and the community site is at community.qdata.tech, with mirrors on Gitee and AtomGit as well as GitHub. The default branch is develop, which tells you the development line and the release line are not the same branch.

Task types accept two engines, which is the real design decision

The feature table is worth reading for the choices it offers rather than the checkboxes it ticks. Data integration supports visual configuration of ingestion, cleansing, transformation and output workflows, and new tasks can use either DataX or Spark as the execution engine, which covers light synchronisation, related-data retrieval, data import, large-scale offline processing and complex computing.

Data development supports task development through SQL scripts, and new tasks can choose Quartz or DolphinScheduler as the scheduler, for data processing, statistical analysis, periodic scheduling and task orchestration. Metadata management has the same choice for collection tasks, so the scheduler is a per task decision rather than a platform wide one.

That is a distinctive design and it cuts both ways. On one side, a team can start a small sync on DataX and a heavy nightly aggregation on Spark, and can move a scheduling job from Quartz to DolphinScheduler without a migration, because the choice lives in the task definition rather than in the deployment. On the other side, every task type now has to be validated against two backends, so the surface area of what can go wrong is larger than a single engine would give you. The third-party dependencies table in the README lists DolphinScheduler as the orchestrator component, so the project expects it in the picture.

The rest of the feature list is conventional for this category. Data modeling covers data standards, warehouse layers, data domains, subject-area planning, logical models and standard data elements. Data quality runs inspections and processing against audit rules, aimed at completeness, uniqueness and validity problems. Data assets gives cataloging, tags, details and search. Data query runs online SQL against data sources for ad-hoc checks and export. Data services packages tables or query results as APIs with online testing and call logs.

The module layout shows how the monolith is divided

The repository top level is where the architecture is visible. There is a qdata-server module for the application itself, a qdata-framework directory for the shared framework code, and a qdata-ui directory for the Vue front end. Then come the business modules, named by abbreviation: qdata-module-da for data assets, qdata-module-dg for data governance, qdata-module-dp for data platform work, qdata-module-dpp for the data platform, qdata-module-ds for data service, qdata-module-dm for data modelling, qdata-module-mc for metadata collection, qdata-module-ai for the AI features, and qdata-module-att and qdata-module-system for the system administration side.

Two more directories are worth calling out. qdata-executor-etl is the ETL execution side, which matches the DataX and Spark choice in the feature table, and qdata-api-ds is a separate API module for the data service, which suggests the published APIs are a deployable unit of their own rather than a controller inside the main server.

The supporting directories complete the picture. sql/ holds the database scripts, docs/ the documentation sources, images/ the assets for the badges and the banner, and docker/ the container setup. The three README files and the mirror badges on Gitee and AtomGit point to a project developed in China and published to several forges at once, which is worth knowing if you are pulling changes and want to know which remote is authoritative.

The backend stack in the README's technology table is a conventional Java stack: Spring Boot for the application, Spring Security for authentication and permission control, MySQL, PostgreSQL, DM8 and KingbaseES for storage and configuration, MyBatis-Plus for the data access, Redis for caching and distributed locks, and RabbitMQ for asynchronous and decoupled processing. The front end is Vue 3 with Element UI and Vite.

Text2SQL is the headline, and the open source edition is scoped

The AI-powered data Q&A module supports natural-language data questions through Text2SQL, with intelligent charts and result detail viewing, and the stated purpose is to lower the barrier for business users who work with data. It is the feature the README leads with in the module table, and it is also the one most clearly positioned as a product differentiator.

The scope of the open source release is stated in a note above the feature table. qData uses a modular design, and the current open source edition focuses on core capabilities including data integration, data development, data modeling, metadata, data quality, data assets, data services and AI-powered data Q&A, with a link to a feature overview page for the details. The wording is careful. It lists what the open source edition covers, and points elsewhere for more, which is the open-core pattern rather than a claim that everything is included.

The two demo environments make the split concrete. There is an open-source edition demo, and a professional edition demo for which you have to contact customer service to get an account. The roadmap section names the areas still to be worked on, with metadata comparison, business layering and data asset refactoring at the front, and continued enhancement of data integration, data quality, data services and AI capabilities behind them, along with more data sources and ETL components.

For an evaluator, the useful question is therefore not whether qData has Text2SQL but whether the open source edition's module set covers your process, because the professional edition is a separate product with a sales process attached.

Demos, documentation and the licence question

The project ships two ways to look before you install anything. There is a documentation community site, and there is an open-source edition demo you can sign into with a published account. A professional edition demo exists as well and requires contacting customer service for credentials. The online documentation link in the README is the community site, and the feature list links into pages hosted there, including the business policy page used for the demo request.

The licence needs a careful reading. The README carries an Apache-2.0 licence badge, and the metadata for the project on GitHub is not classified as one of the standard licences, which points to a custom licence, and there is a LICENSE file at the root of the repository. Where a badge and a licence file disagree, the file is the document that governs, and the README also notes that the open source edition is a foundation for further development, so anyone planning to extend and redistribute it should read the terms rather than rely on the badge.

On maintenance, the project is active. The last push was on 2026-10-20, and the release history shows v1.6.1 on 2026-08-21, v1.6.2 on 2026-09-04 and v1.6.3 on 2026-09-28, with a badge already showing v1.6.4. The default branch is develop rather than master, which matches the develop branch convention in which new work lands and releases are cut from, so a team that pins a tag rather than the branch gets a stable artefact.

Editorial conclusion

qData suits an enterprise team that wants a single platform for data integration, governance and publishing, and that has the Java, Vue and middleware infrastructure to run and extend it, since the repository is a full application rather than a library. It suits a team that needs a small ETL tool or a standalone Text2SQL service less well. Before adopting it, read the LICENSE file at the root, because the README carries an Apache-2.0 badge while the licence GitHub can detect is a custom one, and check the open-source edition's module coverage against the professional edition, since the open source release focuses on the core capabilities. The default branch is develop, the last push was on 2026-10-20, and the newest tagged release is v1.6.3 from 2026-09-28.

Frequently asked questions

Official sources

1. Issues
2. Project website
3. qiantongtech/qData on GitHub
4. README
5. Releases