湛卢智能编码助手

湛卢智能编码助手，融合海量代码库、技术文档、研发知识，为研发人员提供智能、高效的编程支持。可准确理解开发者的意图，将需求转换为高效的代码；已支持AI程序员，具备全栈编程智能体能力，可以快速理解需求、拆解任务，独立规划，自我反思，实现完整编程任务，支持批量单元测试生成、代码重构等项目级代码生成等功能，为企业级代码生成提供智能化解决方案。

兼容 Visual Studio Code、JetBrains IDEs 等主流 IDE；

支持 Java、Python、Go、C/C++/C#、JavaScript、TypeScript、PHP、Ruby、Rust、Scala、Html、 Node.js、Css3、Swift、Dart等主流编程语言。

---

Zhanlu Intelligent Coding Assistant，which integrates massive code repositories, technical documents and R&D knowledge to provide intelligent and efficient programming support for developers. It comprehends the intention of developers accurately, and can transform product requirements and prototype diagrams into efficient code. It supports AI programmers and is equipped with the capabilities of a full-stack programming agent and tool usage. Zhanlu can quickly understand requirements, break down tasks, independently plan, and reflect on its own performance to complete entire programming tasks. It also supports the generation of batch unit tests and code refactoring, among other project-level code generation functions. Additionally, it enables developers to configure MCP tools, ensuring that the coding process integrates seamlessly with their workflow. It provides an intelligent solution for enterprise-level code generation.

Supported IDEs: Visual Studio Code, JetBrains IDEs.

Supported languages: Java, Python, Go, C/C++, JavaScript, TypeScript, PHP, Ruby, Rust, Scala and other programming languages.

核心场景

代码生成

可根据当前代码文件及跨文件的上下文，为您生成行级/函数级代码、单元测试、代码优化建议、图生代码等。支持项目级别workspace问答，帮您梳理项目代码逻辑、生成项目架构；支持代码评审、commit提交信息生成及readme信息生成，辅助您的研发全流程，让您更专注在技术设计，高效完成编码工作。

行/函数级实时续写

在编码的过程中，湛卢可根据上下文和当前语法，自动预测和生成建议代码，按下 Tab 键即可采纳。

自然语言生成代码

在编辑器中，通过自然语言描述你想要的功能，根据你的描述和上下文，支持在编辑器区直接生成代码及相关注释。

研发问答

拥有海量研发文档、产品文档、通用研发知识等，并结合工程级感知能力，为开发者解决编码过程中遇到的研发问题，协助开发者进行代码问题修复、调试或运行错误的排查等。

单元测试生成

湛卢可根据选中的代码片段，结合上下文和跨文件信息，未代码生成单元测试，支持 JUnit、Mockito、Spring Test、unit test、pytest 等框架。

代码优化

湛卢可以深度分析代码，根据上下文及跨文件内容，对代码潜在的编码问题进行解析，从简单的语法错误到复杂的性能瓶颈，帮你快速发现问题，并给出优化建议和优化后的代码。

代码注释生成

湛卢可对你选中的代码快速生成注释内容，支持方法注释和行间注释。

代码翻译

支持一键完成跨语言代码翻译，支持 Java、Python、Go、JavaScript、TypeScript、C/C++、C# 等主流语言，大大节省代码重构时间。

多轮问答

遇到编码疑问、技术难题时，一键唤起湛卢，输入问题描述，湛卢可依据上下文内容和跨文件内容进行多轮问答交互，给出流畅的回复。

互联网检索问答

点击对话框内“互联网问答”按钮， 并输入相关关键词，可以发起互联网检索，湛卢会展示检索结果，使您不离开 IDE 就能浏览需要的信息，实现代码知识实时检索。

知识库问答

基于企业研发管理领域、企业组件库，进行针对性问答，对用户提问给出更符合企业要求的专业答案，构建企业研发知识问答助手，提升团队的工作效率和协作能力。

报错信息一键修复

当编译区出现编译错误或运行出现异常报错时，一键启动报错排查的智能答疑，可结合运行代码、异常堆栈等报错上下文，快速给出排查思路或修复建议代码。

多模态问答

• 支持图生代码，高保真原型图生成前端代码能力，目前已支持element、vue组件库30+组件的图生代码能力。
• 支持图片代码问答，可以识别图片中的代码，进行问答。

AI程序员

AI程序员支持从0到1快速生成项目模板和完整需求代码，实现从理解需求、拆解任务、编写代码、保存执行、修改bug的全流程自主编程，支持基于项目全局视角进行代码优化，也可实现批量单元测试生成、不同编程语言迁移重构等项目级代码生成，提升工作效率。

特色模式

1. 架构师模式：解答项目相关问题，帮您对项目需求做出详细规划。
2. AI程序员模式：您的编程智能体，可根据您的需求自主生成项目级代码，支持文件编辑及MCP调用，实现完整编程任务。
3. 单元测试模式： 支持项目单元测批量生成。
4. 项目缺陷修复模式：基于编程智能体构建缺陷自主修复全流程，输入缺陷一键修复。
5. 漏洞修复模式：上传安全扫描报告，自主对安全漏洞机型核验、分类、修复或生成澄清建议。
6. 代码评审模式：支持项目代码全量评审，并基于评审内容自主优化代码。
7. 技术文档模式：读取项目信息，分析并输出项目概览内容。
8. 智能问答模式：不引入当前文件的上下文，直接进行研发问答。

Core Features

Code Completion

Based on the current or cross-file context, it can generate line-level and function-level code, unit tests, code optimization suggestions, and code from diagrams. It supports project-level workspace Q&A, helping you organize project code logic and generate project architecture. Additionally, it supports code review, commit message generation, and README information generation, assisting you throughout the entire development process. It allows you to focus more on technical designs and complete your coding tasks with greater efficiency.

Line/Function-Level Real-Time Completion

While you code, Zhanlu predicts and suggests the next lines or functions based on context and current syntax; simply press Tab to accept.

Natural-Language-to-Code Generation

Describe the feature you want in plain language right in the editor, and Zhanlu will generate the corresponding code and comments inline, taking your description and surrounding context into account.

Ask Mode

With access to a vast repository of technical documentation, product manuals, and general development knowledge, combined with engineering-level perception capabilities, the Ask mode assists developers in resolving coding issues encountered during development. It helps with code problem fixes, debugging, and troubleshooting runtime errors, ensuring a smoother development process.

Unit-Test Generation

Zhanlu can generate unit tests for any selected code snippet by leveraging full project context (including cross-file information). It supports JUnit, Mockito, Spring Test, unittest, pytest, and other major frameworks.

Code Optimization

By deeply analyzing your code against its surrounding context and the entire codebase, Zhanlu identifies issues ranging from simple syntax errors to complex performance bottlenecks, then provides both optimization suggestions and the refactored code.

Comment Generation

Instantly generate method-level or inline comments for any selected code fragment.

Cross-Language Code Translation

One-click translation between mainstream languages—Java, Python, Go, JavaScript, TypeScript, C/C++, C#, and more—accelerates large-scale refactoring efforts.

Multi-Turn Q&A

When you encounter coding questions or technical hurdles, summon Zhanlu with a single shortcut, describe the problem, and engage in a contextual, multi-turn conversation that references both local and cross-file code.

Web-Search Q&A

Click “Web Q&A” in the chat window and enter keywords; Zhanlu returns curated web results so you can research without leaving the IDE.

Knowledge-Base Q&A

Tailored to your organization’s R&D standards and internal component libraries, this mode delivers answers that align with enterprise requirements, acting as a dedicated team assistant and boosting collaboration.

Error Fix

When compile-time or runtime errors appear, trigger intelligent diagnostics to receive step-by-step guidance or ready-to-paste fixes that incorporate stack traces and surrounding code context.

Multimodal Q&A

• Image-to-Code: Convert high-fidelity UI mock-ups into front-end code with support for 30+ Element-UI and Vue components.

• Image Code Q&A: Upload screenshots containing code and ask questions about it; Zhanlu recognizes the code and responds accordingly.

AI Programmer

The AI Programmer supports rapid generation of project templates and complete demand-related code from scratch, enabling the entire programming process autonomously, from understanding requirements, breaking down tasks, writing code, saving and executing, to debugging. It also supports code optimization from a global project perspective, as well as batch generation of unit tests and language migration and refactoring for project-level code generation, thereby enhancing work efficiency.

Feature Modes

1. Architect Mode: Answers project-related questions and helps you create detailed plans for project requirements.
2. AI Programmer Mode: Your autonomous coding agent that generates project-level code from requirements, supports file editing, and invokes MCP to complete full programming tasks.
3. Unit Test Generation Mode: Generates unit tests in batches for your entire project.
4. Project Bug Fixing Mode: End-to-end autonomous bug-fixing pipeline powered by a coding agent—input a defect and get a one-click fix.
5. Vulnerability Remediation Mode: Upload a security-scan report to let the agent verify, classify, and remediate vulnerabilities automatically or produce clarification suggestions.
6. Code Review Mode: Performs a full-scale review of project code and autonomously refines the code based on the review findings.
7. Technical Documentation Mode: Reads project information, analyzes it, and outputs an overview of the project.
8. Smart Q&A Mode: Delivers R&D Q&A without importing the context of the current file.

应用场景

1、日常编程工作

面向研发人员，提升试试代码补全、智能建议和及时技术问答服务，极大提升编码速度和准确性，优化日常编程工作流程。

2、代码维护与重构

通过对软件项目的深入分析，模型能够自动识别代码中的潜在缺陷，提供针对性的重构建议和性能优化方案，保障软件质量和项目长期可维护性。

3、教育与培训

针对教育界和企业培训需求，湛卢代码大模型通过实例教学和交互式学习环境，主力新手开发者迅速提升编程技能，为技术人才的培养和储备提供有力支持。

Application Scenarios

1、Daily Programming

Tailored for developers, it delivers real-time code completion, intelligent suggestions, and instant technical Q&A, dramatically accelerating coding speed and accuracy while streamlining daily workflows.

2、Code Maintenance & Refactoring

By deeply analyzing software projects, the model automatically pinpoints potential defects and offers targeted refactoring advice and performance-optimization plans, safeguarding software quality and long-term maintainability.

3、Education & Training

Designed for academic institutions and corporate training programs, Zhanlu model empowers novice developers to rapidly advance their skills through hands-on examples and interactive learning environments, providing robust support for cultivating and expanding the technical talent pool.

帮助文档

• 帮助中心

• 开发者社区

FAQ

湛卢支持哪些语言？

湛卢支持C/C++、Java、Python、Go、JavaScript、TypeScript、Rust、PHP、Kotlin、ObjectC等100+主流编程语言。

是否会上传代码？上传哪些代码？

正常使用湛卢过程中需要利用正在编写的文件中的上文以及其他相关文件，来推理出接下来的下文。因此确实会需要通过网络将正在编写的代码文件内容传输给模型推理。我们承诺不使用这部分上传的代码做任何积累、分析、加工等后续动作。

---

Which languages does Zhanlu support?

Zhanlu supports 100+ mainstream programming languages including C/C++, Java, Python, Go, JavaScript, TypeScript, Rust, PHP, Kotlin, ObjectC, etc.

Does Zhanlu upload code? What kind of code is uploaded?

During normal usage of Zhanlu, it needs to utilize the context from the file being written and other related files to infer the subsequent context. Hence, it does require transmitting the content of the currently edited code file over the network for model inference. We commit not to use the uploaded code for any accumulation, analysis, processing, or subsequent actions.

免责声明

在使用湛卢（"我方"）产品及相关服务（以下简称 "本服务"）前，请您务必仔细阅读并理解透彻本《免责声明》。 请您知悉，如果您选择继续使用本服务，意味着您充分知悉并接受以下使用条件：

您知悉并理解，本服务的输出内容及代码，为使用深度合成技术生成的文本信息，我们对其生成内容的准确性、完整性和功能性不做任何保证，并且其生成的内容不代表我们的态度或观点。

您理解并同意，本服务所为您展示的代码只是 "推荐"，若您选择采纳本服务所推荐的代码，应当视为您实际撰写了此代码，您应当是所产生、选择的代码的唯一著作权人。我方不会就本服务所推荐的任何代码承担安全、瑕疵、质量、兼容等任何保证责任，无论是明示或暗示，您有责任确保您生成的代码的安全和质量（无论其是由您完全自主撰写或者是采纳了本服务提出的建议），我们建议您在使用本服务推荐的代码时采取与使用您完全自主编写的代码时相同的预防措施。

您同意并承诺，不会使用本服务进行违反法律的应用开发，如您使用本服务开展特定行业的业务应用（如教育、医疗、银行行业），将同时遵守相关国家规定的用户数据保护法律和内容管理法律。

您确认并同意，我方不会因为本服务或您使用本服务违反上述约定，而需要承担任何责任。

Disclaimer

Before using Zhanlu ("we") products and related services (hereinafter referred to as "this service"), please be sure to carefully read and thoroughly understand this "Disclaimer". Please be aware that if you choose to continue using this service, it means that you are fully aware of and accept the following terms of use:

You acknowledge and understand that the output content and code of this service are text information generated using deep synthesis technology. We make no guarantees regarding the accuracy, completeness, and functionality of the generated content, and the generated content does not represent our attitudes or viewpoints.

You understand and agree that the code displayed by this service is merely a "recommendation". If you choose to adopt the code recommended by this service, it should be considered as if you actually wrote this code, and you should be the sole copyright holder of the generated and selected code. We will not assume any guarantee responsibility for security, defects, quality, compatibility, etc., whether express or implied, regarding any code recommended by this service. You are responsible for ensuring the security and quality of the code you generate (whether it is completely written by yourself or adopts suggestions from this service). We recommend that you take the same precautions when using code recommended by this service as you would when using code you wrote entirely on your own.

You agree and promise not to use this service for application development that violates laws. If you use this service to conduct business applications in specific industries (such as education, medical, banking industries), you will simultaneously comply with relevant national regulations on user data protection laws and content management laws.

You confirm and agree that we will not bear any responsibility due to this service or your use of this service in violation of the above agreements.

致谢

湛卢智能编码助手是基于Roo-Code和Cline开发的，我们深深感谢开源社区的贡献。正是因为站在了巨人的肩膀上，我们才能够打造出这款产品。

我们要特别感谢以下项目对湛卢的重要贡献：

• Roo-Code：其出色的交互设计和丰富的功能特性为湛卢提供了重要的参考和灵感
• Cline：为湛卢提供了强大的VSCode IDE底层能力封装和抽象，是我们构建编辑器扩展的重要基础

这些项目展示了开源协作的变革力量，我们很荣幸能够在它们的基础上继续构建和创新。

Acknowledgments

Zhanlu Intelligent Coding Assistant is developed based on Roo-Code and Cline. We deeply appreciate the contributions of the open-source community. It is precisely because we stand on the shoulders of giants that we have been able to create this product.

We would like to especially thank the following projects for their important contributions to Zhanlu:

• Roo-Code: Its excellent interaction design and rich functional features have provided important reference and inspiration for Zhanlu
• Cline: Provided Zhanlu with powerful VSCode IDE underlying capability encapsulation and abstraction, which is an important foundation for building our editor extension

These projects demonstrate the transformative power of open-source collaboration, and we are honored to continue building and innovating on their foundation.