In May, Qoder CN upgraded the AI Chat (Ask) mode to an Agentic mode, capable of autonomously using project awareness and web search tools, making developer interaction with Qoder CN more aligned with project scenarios. It also added Inline Chat and Next Edit Suggestion (NES) features to make code modification more efficient; and released the AI IDE — Qoder CN IDE, which comprehensively integrates intelligent coding capabilities for out-of-the-box simplicity.
The AI Chat (Ask) mode supports enabling tool invocation. Once enabled, it can autonomously call project awareness tools, such as project retrieval, file retrieval, and code retrieval, without developers manually adding project files. This allows automatic project awareness and answers that better fit the project scenario.
At the same time, when Web tools are enabled, it can autonomously decide to call web search-related tools based on developer questions, such as fetching web page content or querying the network.
For details, see: Ask Mode.
Added the Next Edit Suggestion (NES) feature, which dynamically predicts the next code modification suggestion for developers based on the current code edit. Developers only need to press Tab to efficiently complete code writing. In scenarios such as new feature development, legacy project maintenance, and complex project iteration, it leverages project awareness capabilities to predict edits in the current file as well as cross-file suggestions, quickly completing code modifications within the project.
For details, see: Next Edit Suggestion.
Inline Chat allows developers to have conversations in the code editor area. Developers can modify code within a single file or ask instant questions via natural language dialog.
For details, see: Inline Chat.
Added Qoder CN IDE, which comprehensively integrates intelligent coding assistant capabilities out of the box, allowing developers to enjoy an efficient and intelligent programming experience without installing plugins.
For details, see: Qoder CN IDE Changelog.
In Agent mode, if MCP tools are configured, you can enable auto-execution in settings. When the agent needs to execute the MCP tool, it will do so automatically without waiting for developer confirmation, improving task execution efficiency.
At the same time, the terminal command whitelist has been improved. During a session, if developers want certain commands to be run automatically in the future, they can check the option and click run; the terminal command will be directly added to the whitelist, and the agent will automatically run this command in the future.
AI Chat (Ask) Supports Autonomous Tool Invocation
The AI Chat (Ask) mode supports enabling tool invocation. Once enabled, it can autonomously call project awareness tools, such as project retrieval, file retrieval, and code retrieval, without developers manually adding project files. This allows automatic project awareness and answers that better fit the project scenario.
At the same time, when Web tools are enabled, it can autonomously decide to call web search-related tools based on developer questions, such as fetching web page content or querying the network.
For details, see: Ask Mode.
Added Next Edit Suggestion (NES)
Added the Next Edit Suggestion (NES) feature, which dynamically predicts the next code modification suggestion for developers based on the current code edit. Developers only need to press Tab to efficiently complete code writing. In scenarios such as new feature development, legacy project maintenance, and complex project iteration, it leverages project awareness capabilities to predict edits in the current file as well as cross-file suggestions, quickly completing code modifications within the project.
For details, see: Next Edit Suggestion.
Added Inline Chat
Inline Chat allows developers to have conversations in the code editor area. Developers can modify code within a single file or ask instant questions via natural language dialog.
For details, see: Inline Chat.
Qoder CN AI IDE Released - Qoder CN IDE
Added Qoder CN IDE, which comprehensively integrates intelligent coding assistant capabilities out of the box, allowing developers to enjoy an efficient and intelligent programming experience without installing plugins.
For details, see: Qoder CN IDE Changelog.
MCP and Terminal Commands Support Auto-Execution
In Agent mode, if MCP tools are configured, you can enable auto-execution in settings. When the agent needs to execute the MCP tool, it will do so automatically without waiting for developer confirmation, improving task execution efficiency.
At the same time, the terminal command whitelist has been improved. During a session, if developers want certain commands to be run automatically in the future, they can check the option and click run; the terminal command will be directly added to the whitelist, and the agent will automatically run this command in the future.
Improved Experience for Adding File Directories and Files as Context
- When adding context, you can perform mixed searches across files and file directories, with search highlighting for higher search efficiency.
- Added the ability to drag file directories directly to the input area as context.
Other Features and Improvements
- Agent mode supports selecting file directories (#folder).
- Supports one-click repair when MCP service connections fail.
- Improved large-file modification in AI Chat scenarios.
- Improved knowledge base loading speed when selecting #teamDocs.
- Improved file modification accuracy — when file modification fails, it automatically retries.
- Improved code change generation for Inline Chat.
- Improved pasting of project code snippets into the AI Chat window.
- Improved output when using the qwen3-thinking model in Agent mode.
- JetBrains IDEs now support selecting instructions during a session.