Qoder provides Agentic Chat capabilities in multiple modes: Ask, Agent, Experts, and Quest. These capabilities help developers solve coding problems, fix errors, debug, and troubleshoot runtime issues. Qoder also offers multi-file edits, autonomous decision-making, environment awareness, and tool use to complete end-to-end coding tasks. Install Qoder to explore more.
Qoder provides Agentic Chat capabilities in three modes: Ask, Agent, and Experts. These capabilities help you solve coding problems, fix errors, debug, and troubleshoot runtime issues. Qoder also offers multi-file edits, autonomous decision-making, environment awareness, and tool use to complete end-to-end coding tasks. Install Qoder to explore more.
Core features of Qoder’s Agentic Chat:
- Multiple chat modes
Within a single conversation flow, developers can freely switch between Ask, Agent, Experts, and Quest modes. This flexibility enhances productivity and efficiency across development workflows.
Within a single conversation, you can freely switch between Ask, Agent, and Experts modes to adapt to different development scenarios.
- Automatic environment awareness
Qoder automatically detects project frameworks, technology stacks, required code files, and error messages from task descriptions, eliminating the need to add context manually.
- Tool use
Qoder autonomously uses over 10 built-in tools to assist with file read/write, code queries, and error troubleshooting. It also supports MCP tool configuration, allowing you to customize your toolset as needed.
- Command execution
Qoder autonomously determines, generates, and executes necessary commands, significantly improving task execution efficiency.
- Project-level changes
Based on the task description, Qoder can modify multiple code files within a project. Through multi-turn conversations, it supports code optimization or snapshot rollbacks.
- Memory awareness
Qoder features LLM-based autonomous memory. It learns from each chat, gradually building a rich memory base related to you, your projects, and encountered issues.
Start a new chat
Open the Chat panel
After logging in to Qoder, toggle the secondary side bar in the top-right corner to open the Chat panel.
Alternatively, use the keyboard shortcut:
| Action | macOS | Windows |
|---|
| Open/Close Chat panel | ⌘ L | Ctrl L |
Select a mode
- Ask: A simple Q&A mode that answers coding questions. It gives solutions and suggestions based on context, but doesn’t modify code.
- Agent: An autonomous coding task execution mode that features self-directed decision-making, environment awareness, and tool utilization. Based on the developer’s coding requirements, it leverages project search, plan making, file editing, terminal operations, and other tools to complete coding tasks end-to-end. It also supports developer-configurable MCP tools, ensuring the coding workflow aligns closely with individual development processes.
- Experts Mode: A multi-agent collaboration mode for complex development tasks. Simply describe your requirements, and the Team Lead will automatically break down tasks, assemble an expert team, and execute solution design, coding, testing, and quality review in parallel to deliver production-ready engineering results. Ideal for full-stack development, complex debugging, and technical research.
- Quest Mode: Qoder’s autonomous coding mode. Simply describe your goal, and Quest will autonomously clarify requirements, plan solutions, execute code, and verify results—with minimal human intervention. Ideal for completing development tasks end-to-end.
- Ask: A Q&A mode that answers coding questions. It provides solutions and suggestions based on context, but doesn’t modify code.
- Agent: An autonomous coding mode with self-directed decision-making, environment awareness, and tool utilization. It leverages project search, plan making, file editing, terminal operations, and other tools to complete coding tasks end-to-end. It also supports MCP tool configuration, ensuring the workflow aligns closely with your development process.
- Experts: A multi-agent collaboration mode for complex development tasks. Simply describe your requirements, and the Team Lead will automatically break down tasks, assemble an expert team, and execute solution design, coding, testing, and quality review in parallel to deliver production-ready results. Ideal for full-stack development, complex debugging, and technical research.
After selecting a mode, describe your requirements in the input box. The following tips can help you get better results:
- Structure your request: Clearly state what you want Qoder to accomplish, and outline the goals and steps for the coding task.
- Provide context: Include files, images, code changes, and other relevant information to help Qoder understand the background and generate more accurate solutions.
- Specify expectations: State preferences or guidelines such as programming language, coding standards, output format, or change objectives. Example: “When generating code changes, also include comments for each method.”
- Iterate with feedback: Give feedback on code suggestions or answers. For complex tasks, break down requirements and iterate step by step with Qoder.
Separate chat windows
Separate chat windows let you split an active conversation into its own standalone window. When running multiple Agent tasks simultaneously, you can monitor each task’s progress and results in parallel — no more switching between tabs.
How to use
Right-click a conversation tab at the top of the chat panel and select Move into New Window. The conversation will open in a separate window.
The detached window is fully independent from the main window. You can freely resize and position it, and arrange windows across multiple screens.
Use cases
- Parallel task monitoring: When running multiple Agent tasks, assign different conversations to separate windows to track each task’s status in real time
- Side-by-side comparison: When you need to compare outputs from two conversations, place them in left and right windows for easy reference
Action recommendation cards
After an agent completes a task and responds, a set of recommendation cards automatically appears at the end of the message. Based on the current code changes, project state, and conversation context, these cards predict your most likely next action — click to fill the input box directly, then send or edit further. After the agent finishes execution, new cards are generated based on the latest context.
Action recommendation cards are triggered in Agent Mode and Experts Mode.
Recommendation types
- Review Code Changes: Triggered after the agent generates or modifies code, to review whether the new code meets project standards and check for potential issues
- Run Feature Test: Triggered when frontend files (.tsx, .jsx, .vue, .html, .css) are changed, to verify page rendering and interactions in the browser
- AI Suggestion: Generated by the agent based on the current conversation context
- Recommendations are inferred from context and may not exactly match your intent
- Recommendation cards are only clickable on the latest agent response; they disappear after you send a new message
- Tasks triggered by clicking a card follow the same billing rules as the model selected for the current session
- Click the microphone icon on the right side of the input box
- Speak your requirements — your speech is automatically converted to text
- Review and edit the transcribed text, then send
Tips:
- Voice input supports the same languages as the speech recognition service on your system
- You can combine voice input with manual typing and context attachments
Make plans by To-dos
Based on your input, Qoder generates a plan with a list of To-dos for you to review, breaking down complex problems into manageable, sequential steps.
Code modification and review
Multi-file edits
In Agent Mode, Qoder may modify multiple code files. Each modification involves both a generation and application process. You can view the file statuses in the chat box or workspace:
- Generating: Code suggestions are generated based on task breakdowns.
- Applying: Suggestions are integrated with the original files to create change files.
- Applied: Code changes are completed and ready for review.
Click a file to view the suggestion generation process and a diff comparison of the changes.
Review, accept, or reject modifications
Click the View Changes button to compare modifications. Then:
- Use the up or down arrow to navigate and view changes in the current file.
- Accept or reject each change.
- Use the forward and backward arrows in the file-level operation area to switch between changed files.
- Accept or reject at the file level.
- Partially modify change files.
Multi-turn iterations
Refine requirements in multiple turns
In Agent Mode, after completing a round of conversation and generating code changes, you can continue to refine your requirements by submitting additional queries. Qoder will incorporate the previous code changes, analyze the updated requirements, and produce new code change files accordingly. Qoder also generates multiple snapshots based on your requirements. To revert operations, click the Undo button in the chatflow.
Start a new chat
Click the add button in the upper-right corner of the Chat panel.
View chat history
Click the history icon in the upper-right corner of the Chat panel to see all chat histories.
Context
Qoder supports rich contextual information, including code files, directories, images, Git commits, and rules. You can flexibly combine context with your input to clearly articulate your requirements.
For more information, see Context.
Memory
Qoder offers long-term memory capabilities. As you interact with Qoder, it gradually builds a rich memory base related to you, your projects, and encountered issues. This memory is automatically organized and updated over time, enabling Qoder to better understand your unique needs and preferences.
For more information, see Memory.
Qoder provides various built-in tools that cover common coding operations:
- File searching
- File reading
- Directory reading
- Semantic symbol searching
- File editing
- Error checking
- Command execution
When using these tools, Qoder operates autonomously, making decisions and executing actions without requiring your confirmation or intervention.
For more information, see Tools.
MCP
Qoder’s Agent Mode integrates with MCP (Model Context Protocol) servers. You can configure your own MCP servers for the agent, expanding the AI coding assistant’s capabilities to better align with your workflow.
The agent also connects to a marketplace for third-party MCP servers, allowing you to install the required MCP servers with one click.
For more information, see MCP. 
