Qoder CN JetBrains Plugin Changelog

• Added plugin management to personal settings, with support for custom plugins
• Added conversation forking to explore alternative approaches without affecting the original session
• Added support for including database objects such as tables and schemas in conversations, allowing the model to work with the actual database structure
• Added support for configuring custom models with BYOK
• Added keyword highlighting and filtering to chat history search, making previous conversations easier to find
• Improved sign-in state stability with automatic recovery when authentication is lost, so transient credential failures no longer require a manual sign-in
• Improved agent terminal task execution and the handling of long-running commands

• Fixed an issue where some tool calls in subagents might fail to execute

• Refreshed the visuals and interaction details of the chat interface for a cleaner, more consistent conversation experience

• Added text selection in model responses, allowing selected content from AI responses to be added to the conversation context
• Added a conversation turn navigation pop-up that shows the current and total number of turns and supports quick navigation
• Added a Quick Fix TODO option to the TODO tool window context menu
• Improved background service upgrades and graceful shutdown for smoother plugin updates
• Improved terminal command execution by automatically moving timed-out commands to the background and enhancing WSL workspace support

• Added the built-in Better Harness Skill to evaluate Harness engineering practices and provide targeted improvements
• Improved terminal execution to adapt to the workspace environment; WSL projects now use the IDE terminal by default
• Added a focus-independent PTY mode so terminal commands can run without keeping the IDE window focused
• Added DBML diagram rendering
• Fixed an occasional issue where token-based sessions were disconnected
• Added Explain Code to the editor context menu for selected code (Alt+Shift+P)
• Improved Graphviz rendering compatibility in conversations
• Fixed an occasional credential-reading failure that incorrectly prompted users to sign in again
• Fixed an issue where historical data was lost after configuring a custom local storage path
• Added a full terminal access mode that runs commands without approval
• Added custom agent management for creating, configuring, and invoking custom agents in the plugin
• Added built-in Computer Use and Plan Mode agents, with configurable execution policies
• Added nested subagent presentation, including subtask summaries, collapsible tools, and a task tree
• Added D2, Vega-Lite, Graphviz, QRCode, and other diagram rendering in conversations
• Added LaTeX rendering and improved formula detection to avoid rendering code or plain text as formulas
• Added a sticky prompt summary for navigating long responses
• Added conversation guidance and queue-status notifications, and improved message queue presentation
• Added a gutter icon for assigning TODO comments directly to Qoder
• Added on-demand loading for long conversations when scrolling upward
• Added a bottom-toolbar control for automatically injecting selected text into context
• Added Tools > Qoder > Qoder Cleanup for removing obsolete Qoder files
• Added a dedicated confirmation panel for paused tasks that are ready to resume
• Improved messaging for MCP services disabled by enterprise policy
• Improved stability and loading performance when browsing conversation history
• Improved response rendering for thinking content, Markdown headings, and diagrams
• Improved conversation title management so manually edited titles are not overwritten by AI-generated titles
• Improved AI-generated commit messages for amend operations by including the previous commit changes
• Fixed BYOK credential storage and migration, including automatic migration and clearer recovery guidance
• Fixed an occasional issue where responses remained stuck in the generating state
• Fixed an occasional issue where files could not be edited after reviewing changes
• Fixed DataSpell IDE compatibility issues
• Improved the product experience and fixed known issues

• Added the Qoder first-anniversary campaign
• Added support for enterprise-dedicated model lists
• Added per-conversation model selection memory, restoring the last-used model when reopening a previous conversation
• Fixed issues with BYOK custom model requests
• Improved the usage panel to show plan information only when Credits are available and provide clearer guidance when Credits are exhausted
• Improved the product experience and fixed known issues

• Added migration of local personal AI data after upgrading to Qoder CN
• Added enterprise-configured built-in models, dedicated models, and BYOK providers
• Improved user subscription plan labels
• Fixed an occasional issue where the model selector did not open
• Improved the product experience and fixed known issues

• Added model context and thinking mode configuration
• Improved the personal settings panel and settings presentation
• Improved subscription plan presentation and added Teams support
• Fixed multimodal input switching automatically to the Auto model
• Improved the product experience and fixed known issues

• Improved Credits usage and subscription plan presentation, with support for Pro+, Enterprise Dedicated, and Pro subscription upgrades
• Improved tool-call results and approval information rendering
• Fixed local persistence of the model selector
• Improved the product experience and fixed known issues

• Improved Qoder CN IDE login prompts and remembered the last login entry point

• Improved Qoder CN login guidance

• Improved the product experience and fixed known issues

• Launched the unified Qoder CN personal subscription and Credits system across products
  • One Qoder CN account signs in to the IDE, JetBrains plugin, QoderWork CN, Qoder CN CLI, QoderWake CN, Qoder CN Mobile, and Cloud Agents
  • Existing Credits balances are converted at a 1:1 ratio, with validity recalculated as one month from the upgrade date
  • Upgraded users receive a limited-time benefit of 200 free Qwen3.7-Max calls per day, subject to the official campaign announcement
  • Conversation history, rules, memory, Skills, agents, code indexes, and shared plugin settings are migrated without loss
• Changed the login entry point: upgrade through Alibaba Cloud Account Login, then use Qoder CN Account Login
• Deprecated the legacy Qoder CN VS Code plugin and recommended switching to Qoder CN IDE
• Automatically refunded future subscription-cycle charges through the original payment method and canceled auto-renewal

• Added BYOK model support to prompt enhancement and improved retry behavior after the first request times out
• Added support for IntelliJ IDEA 2025.2+ and running commands directly in the new terminal
• Improved long-conversation compaction with visible status and continued conversation after completion
• Improved the product experience and fixed known issues

• Added multiple tabs to AI conversations for quickly creating, switching, and closing conversations
• Added AMD file separator markers to identify content used as input prompts
• Added Credits usage presentation for enterprise resource packages
• Improved the product experience and fixed known issues

• Fixed known issues

• Fixed an occasional login failure

• Upgraded Lingma to Qoder CN
  • Added support for GLM-5.1 and Kimi-K2.6
  • Migrated Qoder CN (formerly Lingma) subscriptions to Credits-based billing
  • Kept existing enterprise subscriptions on their current pricing and benefits for the contract term; new seats and renewals use the new pricing
• Added one-click prompt enhancement
• Added context-window usage display and manual compaction
• Added Credits usage display

• Fixed an occasional timeout when generating Git commit messages

• Improved the product experience and fixed known issues

• Fixed known issues

• Improved performance

• Added code highlighting for more than 80 programming languages
• Improved performance
• Fixed known issues

• Fixed known issues
• Improved performance and fixed network issues

• Fixed known issues

• Added custom models for Lingma Personal Pro through API keys for Coding Plans from Alibaba Cloud Model Studio, Zhipu AI, Kimi, and MiniMax
• Reduced high CPU usage in some scenarios

• Added custom Skills, available from the input box with /
• Added SkillsJars support for quickly creating a Skill from a Jar package through the Lingma Files menu
• Added shortcuts for starting a new conversation: Alt+N on Windows and Option+N on macOS
• Added files to the Lingma input box from the project file window context menu
• Added cut, copy, and paste operations to the Lingma input box context menu
• Fixed known issues
• Improved performance and fixed network issues

• Fixed known issues

• Fixed known issues

Features

• Introducing the Qwen-Coder-Qoder model:
  • A deeply customized model built to enhance the end-to-end coding experience. Based on Qwen-Coder, it has undergone large-scale reinforcement learning tailored to the Qoder Agent framework, tools, and scenarios. In our real-world software engineering task evaluations, its task resolution rate has surpassed Cursor Composer-1, with terminal command accuracy on Windows exceeding it by 50%.
• Completion and NES fully upgraded to NEXT
  • From passive completion to proactive prediction, with support for automatic dependency import and other features — a comprehensive upgrade in both capability and interaction experience
  • Supports manually triggering NEXT code completion
• Full support for databases via @database
  • Supports referencing databases as context in Lingma Ask/Agent modes
  • Supports generating and executing SQL in the Query Console
• Supports AGENTS.md: when an AGENTS.md file exists at the root of the current project, the agent will automatically recognize and load its content
• Supports adding .aiignore.md to extend restricted file support
• Allows tools to edit files outside the project

Improvements

• Improved several product experience issues
• Supports IntelliJ 2026.1 EAP+ versions

Fixes

• Fixed several known issues

• Fixed some known defects

• Fixed occasional abnormal CPU and memory consumption

• Added the @openFiles annotation to context, allowing all open files—including project files, SDK files, and Scratch files—to be added at once
• Added a tool to enhance real-time awareness of file content in the editor area, improving Agent effectiveness
• Fixed some known defects

• Fixed defects

• Fixed an issue where login state was lost

• Fixed an issue where system paste behavior failed in CLion and Rider
• Fixed an occasional failure of the "Tongyi Lingma one-click fix" feature
• Fixed an occasional issue where the default AI reply language changed to English

• Supports enterprise model selection
• Fixed several defects

• Fixed defects

• Fixed defects

• Fixed an issue where custom instructions failed to execute in specific scenarios
• Improved file editing permission control in the Dedicated Edition
• Fixed several other defects

• Fixed defects

• Coding capability upgrade: Combined with the latest Qwen model capabilities, supporting parallel tool invocation and deeply optimizing coding effectiveness, making complex coding tasks more efficient and accurate;
• New To-Dos tool: The coding agent can break down structured to-dos before starting coding tasks, executing complex long-horizon coding tasks more effectively;
• New message queueing: While Lingma is replying or executing, you can edit and directly send follow-up messages, helping developers iterate quickly and continuously without waiting;
• New fully automatic execution mode: The coding agent supports configurable automatic execution mode. Once enabled, developers can configure a terminal command denylist, greatly improving coding task execution efficiency while ensuring security;
• Enhanced project awareness: Improved retrieval capabilities such as Grep and Search Codebase for more effective project retrieval;
• Enhanced memory awareness: Improved memory and memory organization capabilities so Lingma's replies better understand you;
• Improved tool presentation: The interface presentation of tools such as file read, web retrieval, project retrieval, and directory retrieval has been de-emphasized while confirmation-type actions have been reinforced, allowing developers to quickly grasp key information.

• AI Chat selects Agent mode by default;
• Adding context in AI Chat has been changed to trigger a picker window by typing @ before adding;
• AI Chat (Ask) mode supports retrieval tool invocation by default, and can be turned off in settings;
• After selecting code lines in the code editor area, right-click actions can be added directly to the chat window as context;
• Supports configuring multiple AI rules for a project and configuring how they take effect;
• Improved inline chat editing experience, strengthening the file editing scenario experience;
• Issue reporting supports adding images;
• Removed the model selection entry and file editing mode to simplify developer choices.

• Fixed defects

• Improved the rollback experience for inline code suggestions
• Fixed defects

• Improved inline suggestion prediction generation
• Improved error messages when file modification fails
• Improved error messages for path anomalies
• Fixed occasional plugin installation failures and function-level shortcut failures for Kotlin files in K2 mode
• Fixed IDE lag when executing Gradle clean/build in Android Studio
• Fixed occasional issues where the input box could not receive input

• MCP services support Streamable HTTP
• Improved memory generation
• Improved some UI experiences in MCP services
• Fixed defects

• Code snippets in AI Chat replies support one-click apply to code files
• Code problem fixes (Problem) support Agent mode
• Debug error fix scenarios now respond via Agent mode
• Fixed several defects

• Improved MCP tool parameter configuration experience
• Improved focus prompt style in the input area
• Fixed occasional installation extraction failures for Windows 7 users
• Various other defect fixes

• The MCP marketplace supports English descriptions
• Fixed an issue where language preference was not followed when tools were enabled in AI Chat mode

• AI Chat (Ask) mode supports tools such as project and code retrieval and web search
• Agent mode adds support for web search tools
• Agent mode supports selecting file directories (#folder)
• Adding context supports mixed search of project files and folders with search highlighting
• Improved pasting of project code snippets into the AI Chat window
• Compatibility with the latest IDEs' Reworked Terminal (non-support prompt)
• Supports selecting instructions during a session
• Improved knowledge base loading speed when selecting #teamDocs

• Added Next Edit Suggestion (NES) feature, supporting prediction of the next edit — can be enabled in plugin settings
• MCP tool execution in Agent mode supports configurable auto-execution
• Improved the auto-execution whitelist experience for terminal commands, supporting one-click whitelist addition during sessions
• Supports dragging folders from the file list to the chat window input as context
• Improved file modification accuracy — automatic retry on failure
• Improved pasting of project code into the chat window input
• Improved inline chat code change generation
• Fixed an issue where inline chat could not be seen when it went outside the editor area
• Fixed missing completion count in plugin settings

• Added session mode selection suggestion prompt
• Supports one-click repair when MCP services connect abnormally
• Improved output in Agent mode when using the qwen3-thinking model
• Reduced likelihood of flowchart generation under the code explanation instruction
• Fixed an occasional issue where AI Chat could not detect the currently open file
• Fixed occasional project retrieval output compatibility issues when switching between Agent and File Edit modes
• Fixed occasional file edit mode compatibility issues in PyCharm 2025.1

• Added Inline Chat, supporting Q&A and code modification in the editor area
• Improved large-file modification in AI Chat scenarios

• Tool usage: Autonomously uses more than a dozen built-in coding tools, such as read/write file, code query, and error troubleshooting. Also supports MCP tool configuration and integrates with the ModelScope MCP marketplace, allowing developers to freely configure.
• Terminal command execution: When performing coding tasks, it can autonomously decide which commands to run, automatically writing commands and running the terminal, greatly improving coding task execution efficiency.
• Project-level changes: Based on the developer's task description, it autonomously breaks down tasks and modifies multiple code files in the project. Through multiple dialogs, it supports incremental iteration or snapshot rollback, collaborating with the developer to complete coding tasks.

• Automatic project awareness: Based on the developer's task description, it can automatically perceive the project framework, technology stack, required code files, error information, and other in-project information — no need to manually add project context, making task description easier.
• Memory awareness: Supports LLM-based autonomous memory. During conversations with Lingma, rich memories about the developer, project, issues, and other topics gradually form — the more you use it, the better it understands you.

• Supports navigation to files, methods, and definitions within the project in AI Chat.
• Supports adding project directories when adding context in AI Chat.
• Supports specifying a knowledge base (Enterprise Edition) when adding context in AI Chat.
• Supports multi-select for files, directories, gitCommit, etc. when adding context in AI Chat.

• Supports Alibaba Cloud International account login.
• The AI Chat input box supports Shift + Enter for line breaks.
• Removed the offline completion mode toggle from the status bar.
• Improved the search box focus experience when selecting context.

• Supports developer login for the International Dedicated Edition.

• When entering text in the AI Chat window, related code files can be recommended based on the selected file.
• Supports configuring Project Rules.
• Enhanced Go language effectiveness in inline code completion scenarios.
• Supports one-click issue reporting when startup fails.
• Improved network connection in firewall scenarios.
• Fixed several defects.

• Compatible with JetBrains 2025.1 EAP versions.
• Fixed defects.

• The plugin supports switching between Personal Edition and Enterprise Edition member identities.
• Improved inline completion effectiveness for Java inner class invocation scenarios.
• Improved inline completion similarity awareness.
• Improved automatic context inclusion when selecting input history in the input box.
• Improved full-file awareness when selecting code snippets in AI Chat.
• Improved model reply timeout.
• Fixed several defects in unit test generation in AI Developer.
• Fixed an issue where domain login appeared after logging in for Dedicated Edition scenarios.

• AI Chat and AI Developer support model selection.

• Enhanced AI Developer memory for ongoing session history.
• Mermaid language is now not rendered as a diagram by default, and supports one-click collapse.
• Improved code explanation replies to be more concise.
• Improved apply speed for large-file snippet modifications in AI Developer.
• Improved file path when AI Developer creates new files.
• Fixed file path display issues in generated plans for the unit test agent scenario.
• Fixed an issue where image Q&A failed when questions were asked before image processing completed.

• Supports one-click submission of Qoder CN plugin error logs.
• Improved the experience of resolving startup issues with one-click repair.
• Improved code file generation from images in AI Developer.
• Enhanced SQL generation in code completion scenarios.
• Removed redundant step nodes from the @workspace scenario.
• Fixed several occasional defects in unit test scenarios.

• Enterprise Edition login supports personal token login.
• AI Developer supports generating code change files in the project from images.
• Improved code generation from images in AI Chat.
• Fixed an occasional issue where content could not be copied in unit test scenarios.
• Fixed an issue where multiple files were focused simultaneously in the workspace.

• AI Developer supports batch generation of compilable and runnable unit tests (IntelliJ IDEA).
• Improved inline completion performance.
• Fixed an occasional Diff View error when opened in the workspace.

• Hi, Qoder CN added AI Developer capabilities with multi-file code modification and tool usage capabilities, helping developers complete coding tasks end-to-end, such as feature implementation, issue resolution, unit test case generation, and batch code changes. We also upgraded the developer interaction interface with Qoder CN, adding a unified view for managing changed files and a code change review view. Developers only need to describe requirements in natural language and Qoder CN can modify multiple files and present them for developer review and adoption, greatly improving coding productivity. We hope you enjoy the Qoder CN AI Developer coding experience.

• Continuously enhanced cross-file awareness for Java and Go.

• Improved completion trigger mechanism.
• Improved cross-file completion for Java and Go.
• Enterprise code generation enhancement now supports C/C++.
• Fixed an issue where shortcut keys could not select @ content while replying.

• AI Chat added support for codeChanges and gitCommit context.
• AI Chat multimodal image Q&A supports copy-paste of images.
• AI Chat supports dragging files from the repository directly to the input.
• Simplified AI Chat shortcut instruction names.
• Improved default context inclusion logic in AI Chat.
• Fixed an occasional cursor positioning error in the AI Chat input box.

• Input interaction upgraded to support free combination of context and input text.
• Added support for selecting single or multiple code files as context.
• Added multimodal Q&A mode — after selecting an image, you can perform code generation or issue analysis combining the image content.
• Unit test generation supports awareness of repository POM file configuration.

• AI Chat supports session deletion and copy.
• Fixed defects (various tracking issues).

• Improved local project Q&A effectiveness.
• Fixed various tracking data misalignment issues.

• AI Chat supports Mermaid language preview mode.
• Code completion supports Kotlin cross-file import.
• Code completion strengthens whole-project awareness for Java and Go.
• Improved the in-flight status of the pre-filter security check in AI Chat.
• Improved the copy for enterprise knowledge base Q&A.

• Enhanced library awareness for Python and C#, improving generation.
• Improved personal info page visuals to better match IDE themes.
• Enterprise custom instructions extended IDE compatibility with Windows 7.

• Added support for enterprise custom instructions.
• Improved prompts on login exceptions.
• Improved Go language generation.
• Tried various methods to optimize lag issues.
• Improved startup error prompts.

• Enterprise Edition supports safety filters for AI Chat and code generation.
• Improved rendering of bulleted content in the AI Chat window.
• Improved input box focus interaction after the /clear context command.
• Fixed issues with line-breaking after closing tags and parentheses.

• Optimized library-level awareness and analysis for Go, improving generation.
• Handled empty recall cases in knowledge base Q&A.
• Fixed defects.

• Added Enterprise code generation enhancement, supporting JS/TS and Vue.
• Enhanced project awareness during Java generation.
• Enhanced display of the current login identity.

• Improved page scroll interaction when selecting code with the mouse in AI Chat.
• The plugin settings page supports display language settings.
• Improved inline code generation performance.

• Supports @terminal Q&A for quickly generating commands.
• Improved one-click fix entry interaction for code problems (Problem).
• Fixed an issue where auto-scrolling did not work when selecting code in the AI Chat area.

• Supports setting the display language of the various interfaces in Qoder CN.
• Supports setting the preferred language for AI Chat and commit message generation.
• Improved context awareness for Java code generation.
• Improved cross-file generation for front-end languages.
• Fixed an issue where the inline shortcut toggle for PhpStorm and CLion did not take effect.

• Added support for configuring the visible developer scope by department for enterprise knowledge bases.
• Supports displaying and viewing reference knowledge snippets during enterprise knowledge base Q&A.
• Improved prompts for no-permission and expiration when using enterprise member identity.

• Added inline shortcut instruction entries for Go, Kotlin, Python, Rust, JS/TS, PHP, Ruby, and more.
• Supports enabling/disabling the error troubleshooting entry and improved settings interaction.
• Improved the trigger timing for inline comment completion.
• Improved rendering of AI Chat replies.
• Improved cross-file generation for Java, Go, and C#.
• Enhanced awareness of the current codebase during inline code generation.
• Improved local parsing and indexing performance for Java, Python, Go, and C++ projects.

• Improved the no-permission prompt copy for enterprise member identity.
• Added usage caps for Q&A and inline generation (does not affect usage).
• Fixed startup issues caused by decompression problems and empty info files.

• Supports @workspace for requirement understanding and Q&A based on the current codebase.
• Supports commit message generation.
• The AI Chat window now allows feedback after a downvote — please share more.

• Fixed defects.

• Compatible with the latest JetBrains IDEs 2024.2 EAP versions.

• Supports enterprise private knowledge base Q&A.
• Code blocks in the AI Chat window support direct edit and save after viewing Diff.
• After selecting features such as generate comments, explain code, or generate unit tests, supports entering auxiliary information and reply requirements.
• The input box supports typing while a reply is in progress.

• Supports line-by-line acceptance of inline suggested code.
• Improved rendering of AI Chat replies.
• Improved AI Chat input box experience.
• Improved naming of new unit test files in Python/Go development scenarios.
• Improved bracket matching logic after code completion.

• Improved output smoothness in AI Chat replies — smoother with less jitter.
• Supports modifying the plugin's local storage path.
• Fixed an issue where the AI Chat window could not be triggered in the editor when the Qoder CN icon was missing from the status bar.
• Improved /help content.

• After selecting code, supports triggering generate comments, code optimization, and other actions via shortcuts.
• Compatible with the latest JetBrains IDEs 2024.1 versions.
• Fixed issues such as truncated AI Chat replies and error prompts after login.

• Defect fixes.

• Improved highlighting of code blocks in AI Chat to match the IDE editor.
• Supports auto-focusing the input box when opening the AI Chat window.
• Defect fixes.

• Updated Qoder CN brand assets.

• New unit test files support same-name merge when generated (Java).
• Improved the no-permission prompt during Q&A.

• Supports custom configuration of programming languages that do not trigger inline generation.
• Improved naming rules for unit test files generated in Java development scenarios.
• Improved unit test generation for Java and Python.
• Fixed an error when creating a new file in an AI Chat code block.

• Supports enterprise member identity selection on login.
• Fixed an occasional code parsing failure on startup.
• Fixed several occasional defects.
• Improved inline generation trigger mechanism.

• Added Alibaba Cloud account AK/SK login, addressing Remote SSH, Docker, and WSL usage scenarios.
• Improved code block interaction in AI Chat — can be operated on at any time.
• Added Diff viewing in code optimization suggestions.
• Improved the visual style of the AI Chat input box.
• Improved follow-up logic in AI Chat.

• Added support for viewing historical sessions.
• HTTP Proxy supports the Socks5 protocol.
• Supports configuring whether IDE native completion and inline code suggestions display simultaneously.
• Improved interaction when using code templates while coding.
• Improved text display in Q&A replies.

• Fixed IntelliJ IDEA 2023.3 compatibility issues.

• Added a new feature — Code Optimization, quickly identifying code problems and providing optimization suggestions.
• Supports HTTP Proxy configuration.
• Fixed occasional issues with the auto-update feature.
• Fixed an occasional blank screen issue on the login callback page.

• AI Chat supports including different selected code snippets when asking questions.
• Supports automatic plugin updates with a toggle setting.

• AI Chat supports PHP code highlighting.
• The AI Chat input box supports up/down keys for switching input history.
• Fixed occasional login failures.

• Code blocks uniformly support new file creation.
• Fixed defects.

• Compatible with Windows 7.
• Compatible with Linux arm64.
• Fixed defects.

• Fixed some defects.

• Fixed some defects.

• Line-level/function-level real-time continuation.
• Natural language code generation.
• Unit test generation.
• Code comment generation.
• Code explanation.
• Intelligent error troubleshooting (Java).
• R&D AI Chat.