jeecgboot/jimureport

JimuReport – Open‑source, AI‑enhanced reporting & BI platform

What it is

• A free, web‑based data‑visualisation suite that lets you design reports, dashboards, big‑screen displays, and conversational BI (Chat2BI) without writing code. It ships as a set of Spring Boot starters that can be dropped into any Java backend.

Core capabilities

Area	Highlights
Report Designer	Excel‑like drag‑and‑drop layout, unlimited rows/columns, formulas (SUMIFS, VLOOKUP, IF), multi‑sheet, grouping, cross‑tab, master‑detail, sub‑reports, QR/barcode, conditional formatting, printable templates, data entry forms that write back to the database.
Dashboard / Big‑Screen Designer	Word‑style canvas, 28+ ECharts visual types, map components, video/iframe, auto‑layout, responsive grid, password‑protected sharing, auto‑refresh.
JimuChatBI (Chat2BI)	Conversational interface powered by an LLM (Claude Code or any OpenAI‑compatible model). Users type or speak a request – the AI creates a full‑featured report or dashboard, can refine the result through follow‑up dialogue, and even generate a data‑source schema.
AI‑assisted generation	One‑sentence generation for reports, dashboards, and dashboards; screenshot‑to‑template reconstruction; optional AI‑driven table creation.
Data connectivity	30+ sources: MySQL, Oracle, SQL Server, PostgreSQL, DB2, MariaDB, SQLite, Hive, ClickHouse, Doris, Elasticsearch, MongoDB, Redis, file‑based CSV/Excel/JSON, plus Chinese‑government “信创” databases (达梦, 人大金仓, 神通, 华为高斯, etc.). Supports SQL, API, JSON, WebSocket, and multi‑source joins.
Enterprise features	Fine‑grained printing (invoice, background, batch), data‑entry forms, QR/barcode, mobile‑optimized rendering, role‑based access, scheduled data refresh, custom JS hooks.

AI integration

• The AI assistant is enabled automatically when you use the jimureport‑spring‑boot3‑starter or ‑spring‑boot4‑starter. You only need to supply an OpenAI‑compatible endpoint (DeepSeek, OpenAI, Azure, Qwen, Moonshot, etc.) and an API key in application.yml. The platform then calls the LLM’s chat‑completion API to generate the visual layout and underlying query logic.
• A safety switch (autoTableEnabled) lets the AI create tables in the target datasource – recommended to keep off in production.

How to add it to a project

<!-- For Spring Boot 3 (AI‑enabled) -->
<dependency>
  <groupId>org.jeecgframework.jimureport</groupId>
  <artifactId>jimureport-spring-boot3-starter</artifactId>
  <version>2.5.4</version>
</dependency>
<!-- Optional NoSQL support -->
<dependency>
  <groupId>org.jeecgframework.jimureport</groupId>
  <artifactId>jimureport-nosql-starter3</artifactId>
  <version>2.3.0</version>
</dependency>

Similar coordinates exist for Spring Boot 2 (no AI) and Spring Boot 4.

Running the product

• A ready‑to‑run example (jimureport‑example) is provided; it needs JDK 17+, MySQL 5.7+ (or any supported DB), and optional Redis. Docker images and a “green” no‑install zip are also published.
• After startup, visit http://<host>/login (default admin/123456) to open the web UI, where the AI button appears automatically once the LLM config is present.

License & commercial model

• The code is released under LGPL 3 (dual‑licensing). The free edition can be used commercially as long as the required attribution (“Powered by 积木报表”) remains. Organizations that need a fully proprietary version without attribution can purchase a commercial license from the vendor.

Why it matters for AI‑focused developers

• JimuReport demonstrates a real‑world integration of LLMs into low‑code BI, turning natural‑language prompts into fully‑editable visual analytics. It can serve as a reference implementation for:
  1. Prompt‑to‑SQL generation pipelines.
  2. Automatic UI layout synthesis from textual specifications.
  3. Secure, on‑premise deployment of AI‑augmented enterprise software (the AI connector works with any OpenAI‑compatible endpoint, so you can run a private model).

---

Bottom line – JimuReport is a genuine, open‑source BI/reporting platform that now includes AI‑driven report and dashboard generation. It’s a solid example of how generative AI can be embedded into enterprise data‑visualisation tools.