Qoder CN IDE Changelog

Private Domain Control for Enterprise Dedicated Edition

• Added Private Domain Control (PDC): Qoder CN Enterprise Dedicated Edition consolidates previously scattered internet egress capabilities within the customer's VPC under Security > Private Domain Control, enabling centralized outbound access control, enterprise-managed allowlisting, and end-to-end auditing. Administrators can switch among Isolation, Controlled, and Enhanced modes with one click and fine-tune individual settings. Policies are forcibly distributed to endpoints and cannot be overridden by users. This capability currently supports enablement only for selected enterprise customers. For details, see the Private Domain Control documentation.

Improvements

• Enterprises can now disable public internet access for the built-in browser to meet organization-level security and operations management requirements.

Installation Directory and Main Executable Name Changes

• macOS: The application path changed from Qoder CN.app/Contents/MacOS/Electron to Qoder CN IDE.app/Contents/MacOS/Qoder CN.
• Windows: The installation directory changed from QoderCN to Qoder CN IDE, and the main executable changed from QoderCN.exe to Qoder CN IDE.exe.
• Linux: The package name changed from qoder-cn to qoder-cn-ide, and the main executable path changed from /usr/share/qoder-cn/qoder-cn to /usr/share/qoder-cn-ide/qoder-cn-ide.

App Window Snapshot Now Available on Windows

Features

• App Window Snapshot for Windows: Double-tap Ctrl to capture the frontmost app window, or let real-time voice read the foreground app when you mention what is on screen. The image is passed to the agent as context so it can understand what you are viewing faster.

• Marketplace Chinese Localization: The Marketplace supports a Chinese interface, making plugin browsing, search, and use more user-friendly in Chinese.

Improvements

• Added client information reporting for Qoder CN IDE, helping enterprises centrally manage members' client details.

Qoder CN Application Renamed to Qoder CN IDE

Notice

• macOS: The application path changes from Qoder CN.app/Contents/MacOS/Electron to Qoder CN IDE.app/Contents/MacOS/Qoder CN.
• Windows: The installation directory changes from QoderCN to Qoder CN IDE, and the executable changes from QoderCN.exe to Qoder CN IDE.exe.
• Linux: The package name changes from qoder-cn to qoder-cn-ide, and the executable path changes from /usr/share/qoder-cn/qoder-cn to /usr/share/qoder-cn-ide/qoder-cn-ide.

Improvements

• The one-click prompt improvement button now displays a tooltip by default regardless of input length, making the feature easier to discover and use.
• Knowledge card prefetching is now enabled by default for smoother content loading and reduced wait times.

Conversation-level code review and Spec experience improvements

Features

• Code review now defaults to conversation-level scope: When a review is triggered with the built-in command, it reviews changes related to the current conversation by default.
• Improved Spec experience: You can copy content, quickly open its containing path, and navigate to the downloaded path.

Improvements

• Improved knowledge retrieval in multi-repository workspaces for more efficient cross-repository information discovery.
• Enhanced prompt improvement capabilities, allowing related actions to be triggered more quickly through Slash Commands.
• Hooks now support passing arguments, increasing the flexibility and configurability of automated workflows.
• Improved detection of potentially dangerous commands, covering more high-risk Git subcommands and strengthening pre-execution warnings.

Real-time voice collaboration and scheduled task upgrades | Qoder CN IDE feature alignment with Qoder IDE

Feature updates

• Quest Live Voice supports workspace binding: When starting a real-time voice session, you can select its workspace. The voice task automatically uses the corresponding project context and is archived in the correct location.
• Real-time voice supports voiceprint recognition: Enable voiceprint recognition under Settings > Voice > Live Voice to identify the current user's voice more accurately in multi-person or noisy environments.
• Side tasks: For details or follow-up questions derived from the main task, you can start related side tasks and process them in parallel without interrupting the main workflow.
• Comprehensive scheduled task upgrade: Quickly create tasks through a conversation or manual configuration, with support for more scheduled task types. A new unified management module makes tasks easier and more efficient to view and manage.
• One-click plugin installation through Deeplink: Open a link to start installation, making plugins easier to share and distribute.
• HTTP Hooks: At key points such as tool calls and sessions, events are automatically sent to your configured HTTP endpoint. Based on the response, Qoder can allow or block the action, or inject context, enabling centralized security policies, auditing, and external integrations.

Experience improvements

• Fixed an issue where WSL could occasionally disconnect.
• Optimized the default maximum turn limit for goal-driven execution.
• Improved the responsiveness of the creation flow, reducing wait times when creating tasks and opening related interfaces.
• Improved Search Agent prompts with Markdown structure for clearer, more readable results.

Experience improvements

• Optimized the display of user entitlements in the usage panel.

Fixes

• Fixed known issues.

Qoder Voice real-time voice interaction, Quest plugin marketplace, Better Harness engineering practices, and enterprise capability updates

Feature updates

• Qoder Voice — the first Agentic platform in China to launch real-time voice interaction:
  • Speak to start working: say your idea out loud and it is turned into a task and continuously advanced, saving you the hassle of typing word by word.
  • Cross-screen workflow: the floating ball stays on top of any application, turning what you are viewing into context so Qoder can see your work in progress.
  • Progress like a chat: discuss ideas, organize your thoughts, and clarify requirements through natural conversation—lightweight and without interrupting your flow.
• Quest mode now supports plugin marketplace management: Added details such as plugin categories and download counts, supports setting the installation scope at the user or project level, and provides enable/disable management. It also supports creating and importing custom plugins, helping you discover, manage, and use plugins more efficiently.
• Better Harness is now available for building better harness engineering: Added Better Harness, which inspects harness engineering practices in your project across multiple dimensions and provides targeted improvement suggestions. It supports one-click fixes, with the Agent automatically analyzing issues and generating fix plans.
• Canvas now supports sharing and annotation: Added the ability to share Canvas content within the enterprise, making it easy to share canvas content internally. Also added Canvas annotation, making it convenient to modify and refine canvas content directly.
• Markdown files can now be edited in preview mode: In the Quest window, you can edit Markdown files directly while previewing them, making the operation more convenient.
• Enterprise edition capability updates: The Qoder CN (Full Suite) Enterprise Standard Edition / Enterprise Dedicated Edition also add the following features:
  • Enterprise-level unified MCP control: enterprise administrators can centrally control member access to MCP Servers to protect enterprise data security.
  • Enterprise private plugin marketplace: supports browsing, installing, and using enterprise private plugins, enabling internal plugin sharing and distribution.
  • The Qoder CN (Full Suite) Enterprise Dedicated Edition now supports QoderWake CN and Qoder CN Mobile.

Experience improvements

• Clearer conversation revert: improved the interaction feedback after Revert is triggered, helping you more intuitively understand the revert scope and current state.

Fixes

• Fixed an intermittent timeout issue during terminal execution.

Qoder Security officially released, with new action recommendation cards

Feature updates

• Qoder Security officially released: Focused on security in the AI coding era. Supports L1 / L2 / L3 progressive three-tier scanning, embedding security review into every stage of code generation, conversation, and commit, with one-click quick fixes—so every line of code is secure the moment it is committed.
• New action recommendation cards: Intelligently suggests next-step actions based on the current usage scenario, improving operational efficiency and the overall experience.

Experience optimizations

• Optimized the Qoder CN (formerly Lingma) upgrade experience.

New Enterprise edition, browser visual annotation, and comprehensive Hook upgrades

Feature updates

• Comprehensive Hook capability upgrade: Added five new event types: SessionStart, SessionEnd, SubagentStart, SubagentStop, and Notification. Introduced Hook asynchronous rewake capability (asyncRewake). Tool events now support if conditions to match specific tools.
• Browser visual annotation: Directly select elements and annotate regions during browser operations to assist page visual debugging and adjustment.
• Voice input: After speech-to-text conversion, spoken language is automatically optimized for clearer, more natural input—improving dialog quality with the agent.
• Workspace supports multiple folders: Quest now lets you add multiple folders to a single Workspace, enabling flexible organization and management of multi-directory content in one workspace.
• Support for Qoder CN Series Enterprise subscription: Qoder CN Series Enterprise edition includes all Teams features and adds group-based permission and billing management, multi-dimensional model policy governance, plugin sharing and distribution governance, an enterprise private capability marketplace, operation auditing, and priority service support. Migration from Qoder CN (formerly Lingma) Enterprise Standard Edition to Qoder CN Series Teams and Enterprise editions is supported. After the upgrade, enterprise members can log on with a new account and migrate AI data (such as conversation history, rules, memory, agents, and code indexes) from the original instance account to the new account, preserving history.

Optimizations

• Optimized the Experts task dispatch strategy to avoid duplicate dispatch of Experts with the same name and reduce conflicts between running tasks.
• Improved knowledge card generation and retrieval efficiency for a better organization and lookup experience.

Long-horizon task orchestration, conversation-level Fork, and a fully upgraded Quest window experience

Model context and thinking mode configuration:

Goal-driven and scheduled execution for more reliable long-horizon tasks:

• Goal-driven task execution: Supports setting a goal in the Quest window; the agent will continue executing around the goal until the task is complete.
• One-time scheduled tasks: Supports setting a Goal or Spec as a one-time scheduled task, launching execution at a specified time.
• Set Spec as goal for execution: After generating a Spec, you can directly set it as a goal and initiate execution.
• Steer: Supplement direction or refine requirements at any time during task execution, reducing interruptions and repeated restarts.

Quest window supports mobile remote control:

Conversation-level Fork for flexible branched follow-ups:

Full upgrade to the core Quest experience:

• Task sidebar upgrade: Added grouping, sorting, and show/hide customization components, improving efficiency when viewing and switching between multiple tasks.
• Worktree collaboration: Added uncommitted change handling and worktree cleanup capabilities, making task branch management clearer and more controllable. Optimized the collaboration flow with support for handoff.
• Browser state reuse: Reuse browser state when switching sessions within the same workspace, reducing repeated openings and reloads.
• HTML file preview: Supports previewing HTML files directly in Quest scenarios, making it more convenient to view page effects.
• Quest annotations can be added directly to context: Supports adding annotations to specific content in Spec, code, and Diff views, and delivering the feedback directly to the agent to help it accurately understand the modification requirements.
• File operations and interaction: Enriched right-click menu operations for the file tree, centralizing commonly used file operations. Supports adding external files to the file tree via drag-and-drop and other methods.
• File tree styling: Optimized the overall file tree styling to align with the editor area style.
• Text file handling: Optimized how manually imported text files are handled, avoiding content truncation and misunderstanding.
• Knowledge card export: Optimized the export structure of knowledge cards, improving the clarity and usability of export results.
• Canvas recognition: Optimized Canvas's recognition of project configurations, improving code understanding and diagnostic accuracy.

AI conversation experience improvements:

• Improved the interaction when the input box exceeds the character limit; the excess content will be automatically converted into a file attachment.
• Conversation streams now support more efficient collapsing of lengthy conversation streams.
• Optimized the compression strategy when triggering sync, reducing extra overhead and improving processing efficiency.
• Optimized the wait time for the first message sent in a new conversation, improving initial interaction response speed.

Memory system self-evolution:

Performance and stability:

• Desktop architecture: Improved compatibility and runtime stability.
• Remote SSH network: Optimized the network proxy path, improving transmission stability in remote connection scenarios.

Feature Updates

• Added Qoder CN Enterprise Edition, which supports logging in to Qoder CN IDE, Qoder CN CLI, and QoderWork CN with a single account.
• The Expert Team now supports Qwen 3.7-Max and GLM-5.2.
• Computer Use now supports Windows: Computer Use capabilities are now available on Windows devices.
• More controllable RepoWiki generation: RepoWiki generation now supports upfront intervention. Before generating a RepoWiki, you can first specify the analysis scope and focus perspective, generate a knowledge generation plan file through interactive dialogue to guide RepoWiki generation, and support human-machine collaborative knowledge generation, editing, and rewriting.
• Quest now supports workspace-free sessions: You can start a session directly without opening a workspace first, ideal for quick questions and lightweight tasks.
• Quest tasks now support quick conversation navigation: The Quest window adds a side navigation capability, making it easy to quickly jump to a target conversation segment.

Experience Improvements

• Improved the Qoder CN IDE login prompt to automatically remember the last login entry point.

Experience Improvements

• Improved the Qoder CN IDE login onboarding.

Experience Improvements

• Improved the model selector by adding model descriptions.
• Improved the page prompts for upgrading a personal subscription to Qoder CN (all-in-one bundle).

One Subscription, Universal Across All Products: The Unified Credits System for Qoder CN Personal Edition Is Now Officially Available

New

• One universal account for Qoder CN (all-in-one bundle): A single Qoder CN account seamlessly logs in to the IDE / JetBrains Plugin, QoderWork CN (desktop), Qoder CN CLI (terminal), QoderWake CN (digital employee), Qoder CN Mobile (mobile), and Cloud Agents (cloud Agent), with no need to register separately.
• Unified Credits resource pool: The Credits balance of your original account will be converted at a 1:1 equivalent to your new all-in-one bundle account, with the validity period recalculated as one month starting from the upgrade date.
• Limited-time model benefit: Upgraded users can enjoy a limited-time offer of 200 free daily calls to Qwen3.7-Max (the end date is subject to the official website announcement).
• Full AI data retention: Historical conversation records, personalized rules, memory, skills, agents, code indexes, and common plugin settings are all migrated losslessly to the new account.

Changes

• Login entry point switch: Upgrade automatically through the "Alibaba Cloud account login" entry point. After a successful upgrade, please log in using the "Qoder CN account login" entry point; the original login method will no longer be available.
• VSC plugin support discontinued: The original Qoder CN (formerly Lingma) VSC plugin has been marked as Deprecated and is no longer part of the all-in-one bundle product matrix. We recommend switching to Qoder CN IDE.
• Automatic order refund: After the upgrade, fees for the next cycle and beyond of your original subscription will be automatically refunded via the original payment path, and auto-renewal orders will be canceled accordingly—no manual application required.

Other Features

• Built-in expert agents now support custom prompts: Built-in expert agents now support configuring custom prompts, making expert capabilities better aligned with team standards and personal usage habits.
• Custom sub-agents now support model configuration: When creating a custom sub-agent, you can specify the model as needed, making it easy to choose the model capabilities best suited to different tasks.

Feature Updates

• Quest window now supports custom layouts and the "My Quests" task board: The Quest window supports dragging to split panes for handling multiple tasks in parallel, and adds a "My Quests" board that summarizes all Quests by status, giving you a full view of progress at a glance.
• Expert Team now supports custom models: Flexibly configure the models used by the Expert Team and sub-agents, switching between built-in models or using custom models to adapt to more task scenarios.
• Computer Use and window snapshot capabilities launched on Mac: Once authorized, the Agent can invoke computer operations on macOS and control apps in the background to complete complex automation tasks; double-tapping ⌘ Command captures a screenshot of the frontmost window as context for the agent, enabling more precise understanding (macOS only).
• Support for using Dev Containers in SSH Remote: Dev Container support has been added in SSH Remote scenarios, along with support for attaching to any Docker container, improving the flexibility and consistency of remote development environments.
• Enhanced code review capabilities: Added support for /ultra review, improving the analytical depth and suggestion quality of code reviews, helping you discover potential issues and improve implementations faster.

Fixes

• Fixed known issues

Feature Updates

• Added /canvas visualization generation capability: Supports triggering the generation of interactive Canvas views via /canvas, making content organization, report presentation, and analysis more intuitive.
• Enhanced code review capabilities: Added support for /ultra-review, improving the analytical depth and suggestion quality of code reviews, helping you discover potential issues and improve implementations faster.
• Added resource packs for the enterprise standard subscription of the Credits billing system. Enterprise administrators can also configure resource pack usage limits for members.

Fixes

• Fixed known issues

Fixes

• Fixed known issues

Fixes

• Fixed issues with credits usage display and refresh

Lingma Upgraded to Qoder CN

Feature Updates

• Standalone Quest window & fully revamped Editor: Quest has been upgraded from an in-IDE mode to a standalone window, becoming the command center of the Agent First working paradigm. It supports running Agent tasks for different projects across multiple Workspaces simultaneously, tracking global activity in real time from a unified panel, and automatically generating a delivery checklist when tasks are complete; the Editor window has been fully upgraded, delivering a smoother human-machine collaborative coding experience.
• Experts (Expert Team) mode now officially available: Multiple experts work in parallel for high-quality delivery, covering the full chain from solution design to coding to quality inspection, and supporting an end-to-end pipeline from solution design to code delivery; a self-evolving AI team that understands you better the more you use it, continuously learning your coding style and accumulating dedicated collaboration experience; five types of experts—planning, research, coding, review, and testing—have officially joined Quest, completing end-to-end delivery in a pipeline manner; support for custom experts—configure domain knowledge, skills, and tools to build a dedicated Agent team tailored to your business scenarios.
• Added support for models such as GLM-5.1 and Kimi-K2.6: Providing more model choices for intelligent Q&A and agent mode.

Billing and Entitlement Changes

• Qoder CN (formerly Lingma) has been upgraded to a Credits-based billing system, where all version subscription plans deduct Credits based on actual usage. For details, please refer to Billing Description Billing Description.
• Existing enterprise edition subscriptions will continue at the original pricing and entitlements within the current contract cycle; new seats or renewals will follow the new pricing.

Feature Updates

• Support for upgrading Lingma IDE to Qoder CN IDE:
  • Upgrade to Qoder CN to unlock a brand-new experience. Standalone Quest window: immersive conversations for a more focused workflow. Fully revamped Editor: a smoother human–AI collaborative coding experience. Credits subscribers also gain Experts mode and models such as GLM-5.1 and Kimi-K2.6.
  • Seamless migration, no worries. Your historical conversations, rules, memory, agents, and code indexes are automatically retained.

Fixes

• Fixed known defects

Feature Updates

• Code Review agent: Supports /code-review, with a built-in Code Review agent that can trigger a code review within a specified scope in a conversation while in agent mode
• Browser agent: Supports /browser, with a built-in Browser agent that accesses web pages and reads online documents via the /browser command in agent mode
• Hook mechanism: A brand-new Hook mechanism has been introduced into the Agent execution flow, allowing custom logic to be inserted at key points of agent execution. It supports events such as UserPromptSubmit (user submission), PreToolUse (before tool invocation), PostToolUse (after tool invocation), PostToolUseFailure (tool invocation failure), and Stop (stop)

Experience Improvements

• Clearer conversation rollback operations: Improved the interaction feedback after Revert is triggered, helping users more intuitively understand the rollback scope and current state

Fixes

• Fixed known issues

Feature Updates

• DevContainer support: Added development container (DevContainer) capabilities, supporting the definition of standardized development environments via the .devcontainer/devcontainer.json configuration file, helping teams achieve a consistent development experience
• Agent mode supports Windows terminal sandbox: Commands executed by the Agent can now run in a secure sandbox, effectively isolating potentially risky operations

Experience Improvements

• Extension Host process isolation: Improved the process isolation of built-in AI coding-related extensions by running AI extensions in a separate Extension Host process, enhancing runtime stability and the overall user experience

Feature Updates

• Agent sessions can be opened in a standalone window: Support for opening Agent sessions in a new window, making it more efficient to handle tasks in multi-session parallel scenarios and further improving the user experience
• Quest mode supports re-editing sent messages: In Quest mode, you can now re-edit messages that have already been sent, making it easy to adjust conversation content at any time and improving interaction flexibility
• Qoder CN IDE adds AKSK authentication: Supports logging in via Access Key/Secret Key
• The Spec/plan preview panel adds a file search feature, supporting quick location of target files

Feature Updates

• Skill UI dynamic interface generation capability launched: When the Agent executes a Skill, it can render interactive HTML components in real time within the conversation, supporting rich interactive scenarios such as forms, charts, and configuration panels (on first use, you need to have the Agent create an interface for the corresponding Skill)

Experience Improvements

• Quest mode artifact area navigation upgrade: Improved the artifact area navigation bar, file tree hierarchy, and interaction experience
• Improved the Quest mode experience when no project is open, making it easier to find historical tasks
• Improved Repowiki error code prompts

Fixes

• Fixed defects related to issue reporting

Feature Updates

• Files in the Quest mode artifact area support fine-grained modification via Add to chat—select part of the content and add it to the conversation; this applies to both code files and Spec files

Experience Improvements

• Improved Quest mode task list sorting, arranged in reverse chronological order by most recent conversation
• Qoder CN Enterprise Edition improved the network connectivity of the IDE plugin marketplace

Experience Improvements

• Added a BYOK entry point for configuring third-party models in the model selector

Feature Updates

• Support for custom models: Qoder CN Personal Professional Edition supports connecting to the Coding Plans of Alibaba Cloud Bailian, Zhipu, Kimi, and MiniMax providers via API keys to use third-party models
• Agent mode supports Workspace search and file editing across different projects
• VS Code kernel upgraded to 1.106.3
• Repo wiki upgrade:
• Core experience upgrade: Visualized generation progress, improved operation guidance to distinguish between generation and updating, added import/export instructions for bidirectional sync, and added automatic export capability
• More efficient traceability: Automatically displays the Commit ID after generation for Git-associated projects, and reference source annotations now include line numbers
• New user activity display: Shows the number of remaining free generations (note: Repo Wiki is currently open for Beta experience for the enterprise edition)

Experience Improvements

• The input box now allows adding up to 20 images and 20 files
• Quest mode file tree supports right-click operations: add to context, copy relative path, open in folder
• Improved the empty-project conversation interaction in agent mode

Feature Updates

• Skills and custom agents support visual management: Provides a visual management interface for skills, custom agents, and commands, supporting management operations such as automatic creation and import

Experience Improvements

• Expanded custom agent configuration options, supporting configuring skills for them
• Improved the experience of opening code files in the Quest mode artifact area

Feature Updates

• Built-in create-skill and create-agent skills to guide and simplify creating skills and custom Agents for users. Also supports quickly invoking and using skills and subagents via /
• Added a file tree to Quest mode, improving the ability to view files in the current project and conveniently add them to context
• Quest mode supports invoking the Vercel Skill through conversation to quickly deploy and launch applications

Experience Improvements

• One-click prompt optimization: Improved the handling of @context scenarios contained within prompts
• Support for the .agents/skills directory: Now supports reading skills from .agents/skills, following the unified industry standard
• Custom Subagent execution details support expanding a popup for viewing
• NEXT supports reading clipboard history
• Enhanced the Agent's ability to invoke Skills in Quest tasks
• Improved the experience of creating new tasks and collapsing the toolbar on the Quest mode page
• Updated the example content on the Quest mode home page
• Login stability improvements: Improved login stability (if you encounter login or permission issues after upgrading, please try logging out and logging in again)

Fixes

• Fixed occasional system exception issues
• Fixed occasional file editing freeze issues
• Fixed an issue where context with images could cause the IDE to crash

Experience Improvements

• Improved the handling logic in intelligent Q&A (Ask) mode when tools are not enabled

Feature Updates

• Major model upgrade Qwen-Coder-Qoder: A deeply customized model built to improve the end-to-end programming experience. This model is based on Qwen-Coder and has undergone large-scale reinforcement learning tailored to the Agent framework, tools, and scenarios. In our evaluations of real-world software engineering tasks, its task resolution rate performs excellently, especially with a 50% higher accuracy for terminal commands on Windows systems
• Qoder CN IDE supports Quest autonomous programming, now open for Beta experience:
• Autonomous programming agent: The Agent autonomously completes high-quality, end-to-end, deliverable results without continuous human intervention
• Requirement alignment mechanism: Intent recognition, requirement clarification, and co-created Specs to clearly define the task before execution
• Long-horizon task capability: Greatly enhanced ability to run continuously over extended periods, with the Agent supervising tasks until completion
• Autonomous quality assurance: Integrated result verification capabilities to autonomously verify and fix the quality of deliverables
• Continuous autonomous evolution: Not just executing, but proactively evolving. Remembers your style and continuously learns new technologies
• Support for more application scenarios: Quest performs intent recognition on user input and automatically routes to the most suitable capability
• Spec-driven development: First align requirements and constraints, then execute and accept, ensuring clear and traceable results
• From idea to product: Supports creating websites and prototypes from 0 to 1, integrating top-tier visual models to greatly improve design quality and usability
• Agentic Chat: Multi-agent parallel execution, with a comprehensive upgrade to the built-in planning agent and custom extension capabilities:
• Added a built-in planning agent: When handling complex tasks, users can develop an actionable execution plan in advance through a human-machine collaboration mode. This can significantly improve task execution efficiency, avoid repeated rework, and enhance the quality of Agent-generated code as well as the high controllability of results
• Custom extension capabilities upgraded again: Supports custom agents (subagents), skills (Skills), and commands (Commands) to build your own intelligent workflow
• Multi-agent parallel sessions: Supports running multiple session windows in parallel to handle different tasks simultaneously, significantly improving multitasking efficiency