dataease/dataease: a self hosted drag and drop BI tool with an offline installer, two release lines and a license that keeps the logo

Charts by drag and drop over SQL engines, files and APIs

DataEase is a BI tool built around one interaction: connect a data source, then build charts by clicking and dragging fields, and share the result with others. Its pitch is that people without SQL skills can analyze business trends themselves, and it lists four claims for itself: open and free to obtain with monthly iterations, easy to use through mouse clicks and drag and drop, support for many platforms and embedding scenarios, and several sharing modes meant to keep data secure.

The supported sources fall into five groups. OLTP databases include MySQL, Oracle, SQL Server, PostgreSQL, MariaDB, Db2, TiDB and MongoDB-BI. OLAP databases include ClickHouse, Apache Doris, Apache Impala and StarRocks. Amazon RedShift represents data warehouses and data lakes. Excel and CSV cover file uploads, and there is an API data source type for HTTP endpoints. Each group ends with an et cetera, so the list is not closed, but if your warehouse is not named, confirm it in the online documentation before planning around it. JDBC drivers ship inside the image under a drivers directory, which is where support for a given database actually lives.

The offline installer wants a 2 core, 4 GB Linux server and ships a known password

The supported path is an offline installer. You prepare a Linux server with at least 2 CPU cores and 4 GB of memory, download the offline package from the FIT2CLOUD open source community downloads page, unpack it and run the install script:

tar zxvf dataease-offline-installer-v3.0.0.tar.gz
cd dataease-offline-installer-v3.0.0
bash install.sh

The file name carries the version, so the archive you download for a newer release will not be named v3.0.0. After installation the web console accepts the username admin with the password DataEase@123456. That default is published for everyone to read, so any instance reachable from a network should have the password changed at first login, before you connect production databases. An alternative is the 1Panel app store, FIT2CLOUD's Linux server management panel, which deploys DataEase as an app; the linked guide for that path sits under the v2 documentation tree on dataease.io, which matters for the version question below.

Inside the container: one Spring Boot jar listening on port 8100

The repository's Dockerfile shows how the server is packaged. The base image is alpine-openjdk21-jre, so the backend runs on a Java 21 runtime. The build copies drivers/ into /opt/dataease2.0/drivers/, staticResource/ into /opt/dataease2.0/data/static-resource/ and mapFiles into /opt/dataease2.0/data/map-origin, then adds a single jar, core/core-backend/target/CoreApplication.jar, as /opt/apps/app.jar.

Runtime settings come from environment variables. RUNNING_PORT is 8100. JAVA_OPTIONS sets UTF-8 file encoding, a loader path of /opt/apps and an additional Spring config location of /opt/apps/config/, which is the directory to mount if you want to override configuration without rebuilding. A health check runs nc -zv against 127.0.0.1 on the running port every 15 seconds with a 5 second timeout, 20 retries and a 30 second start period, and the container starts through /deployments/run-java.sh. The health check only confirms the port accepts connections; it does not prove the application can reach its MySQL metadata database, so monitor that separately.

Building the image needs a compiled jar and access to Alibaba Cloud and fit2cloud.com

The Dockerfile does not compile anything. It expects core/core-backend/target/CoreApplication.jar to exist already, so you have to run the Maven build defined by the top level pom.xml before docker build, or the ADD step fails.

Two network dependencies are baked in. The base image is pulled from registry.cn-qingdao.aliyuncs.com, an Alibaba Cloud registry in Qingdao, and a RUN step downloads a blocked.certs file from resource.fit2cloud.com, under a path named for CVE-2026-91403, into the JVM's security directory. For teams outside China the registry pull can be slow, and for air gapped build environments both steps break outright. You would need to mirror the base image and the certificate file internally and edit the Dockerfile, which means carrying a local patch against every upstream release. Teams that cannot do that are better served by the offline installer, which is designed for servers without direct internet access.

Two release lines, v2.10.x and v3.2.x, from a default branch named dev-v2

DataEase ships two major versions at once. v3.2.0 and v3.2.1 were both published on 2026-10-22, and v2.10.28 followed on 2026-10-29. The repository's default branch is dev-v2, and the last push landed on 2026-11-02. Meanwhile the install example names a v3.0.0 package, the license section speaks specifically about DataEase V3, the 1Panel guide lives under docs/v2 on dataease.io, and the container paths still say /opt/dataease2.0.

The consequence is that you must pick a line deliberately. Documentation pages, forum answers and screenshots can refer to either version, and the branch you land on when browsing GitHub is the v2 development branch, not v3. Before following any guide, check which version it targets, and before upgrading an existing installation, read the release notes for the specific tag rather than assuming v2 to v3 is a routine step. Monthly iteration is part of the project's promise, so plan for regular upgrades on whichever line you choose.

The FIT2CLOUD license is GPLv3 with a logo clause

GitHub reports the license as NOASSERTION, and the badge at the top of the README links to GPLv3, but the actual terms are the FIT2CLOUD Open Source License. DataEase V3 is released under it, and FIT2CLOUD describes it as essentially GPLv3 with additional restrictions. You may build on the source code, with two conditions: you may not replace or modify the DataEase logo and copyright information, and derivative works must meet GPLv3's open source obligations.

This rules out a common use of open source BI tools: embedding dashboards inside your own product under your own brand. If you need white labeling, or you do not want to publish your modifications under GPLv3, the route is a commercial license from FIT2CLOUD via [email protected]. Automated license scanners that rely on GitHub's detection will not flag any of this, so have someone read the LICENSE file before DataEase goes into a customer facing deployment.

Calcite, SeaTunnel and Univer under the Vue and Spring Boot shell

The stack is spelled out. The front end uses Vue.js with the Element component library, and charts are drawn with AntV. The backend is Spring Boot, with MySQL as the application database. Data processing relies on Apache Calcite and Apache SeaTunnel, and spreadsheet features come from Univer. Docker is the infrastructure layer.

For operators this explains the moving parts you will troubleshoot: a JVM service, a MySQL database for DataEase's own state, and connectors to each source. For contributors, the top level layout splits into core/ for the main application, drivers/ for database drivers, sdk/, installer/, mapFiles/ for map data, staticResource/, tests/ and docs/. The repository also carries an AGENTS.md and a .codex/ directory for AI coding agents. FIT2CLOUD's related SQLBot project, which answers questions over data using large models and RAG, lives in its own repository rather than inside DataEase.

Chinese first documentation and community

The main README is in Chinese, with translations in docs/ for English, Traditional Chinese, Japanese, Brazilian Portuguese, Arabic, German, Spanish, French, Korean, Indonesian and Turkish. The deeper resources follow the same pattern: online documentation at docs.fit2cloud.com, a community forum at bbs.fit2cloud.com, an introduction video on Bilibili, and an official slide deck for presenting DataEase to a team. A Gitee mirror is linked alongside GitHub. Non Chinese speaking teams can run DataEase, but should expect to read some answers through translation and to search the Chinese forum when the English README runs out.

Editorial conclusion

DataEase suits teams that want a self hosted, drag and drop dashboard tool over common SQL databases, OLAP engines and spreadsheets, and that are comfortable with FIT2CLOUD's Chinese first documentation and community. It is the wrong choice if you need to rebrand the product for customers, because the FIT2CLOUD Open Source License forbids replacing the logo and copyright notice without a commercial license. Before installing, decide between the v2.10.x and v3.2.x lines, confirm your database appears in the supported source list, and change the default admin password immediately.

Frequently asked questions

Official sources

1. Official documentation
2. Official README
3. Project repository
4. Release notes