Overview
Quest Mode is Qoder’s autonomous coding feature, enabling agents to complete development tasks end-to-end. Simply describe your goal, and Quest will autonomously clarify requirements, plan solutions, execute code, and verify results—with minimal human intervention. Core Philosophy: Define the goal. Review the result.Core Features
Autonomous Coding
Agents autonomously deliver high-quality, end-to-end, production-ready results with minimal human intervention:- Powered by leading models: Uses cutting-edge AI models to deliver optimal results right away
- Requirement alignment: Recognizes intent, clarifies requirements, and co-creates specs—ensuring alignment before execution
- Long-running capability: Enhanced long-running task execution—agents monitor and complete tasks autonomously
- Quality assurance: Built-in verification automatically validates and fixes deliverables
Continuous Self-Evolution
Goes beyond execution—learns and evolves:- Remembers your coding style and project patterns
- Continuously learns cutting-edge techniques
- Gets better over time
Support for More Use Cases
Quest recognizes user intent and automatically routes to the right capabilities:- Spec-driven development: Aligns on requirements and constraints upfront, then executes and verifies for clear, traceable results
- From idea to product: Creates websites and prototypes from scratch, with top-tier visual models for exceptional design quality and usability
Long-Running & Parallel
Run tasks locally and remotely in parallel for hours. Self-healing with automatic verification:- Local + Remote parallel execution: Double your efficiency
- Hours-long task support: Handles interruptions and recovery automatically
- Elastic Sandbox: Supports tens of thousands of concurrent tasks
Skills Extension
Extend Quest’s capabilities through Skills:- To use Skills, simply install Skills to
{project}/.qoder/skillsin Quest 1.0
Getting Started
Switch to Quest Mode
Qoder offers two working modes:- Editor Mode: Real-time collaborative coding, conversational
- Quest Mode: Task delegation, autonomous execution, hands-off delivery
- Open Settings > Quest Mode
- Configure Quest Mode Layout option
- Choose which mode to use by default when opening projects
Create a Task
Entry Point
Click the New Quest button at the top of the task list on the leftChoose a Scenario
Quest offers three scenarios—choose based on your needs:| Scenario | Use Case | Quest Behavior |
|---|---|---|
| Code with Spec | Complex features, refactoring, strict quality control | Aligns on scope first, designs implementation and acceptance criteria, then executes |
| Build a Website | 0-1 website creation, rapid prototyping | Describe the website you want to build. Quest will set up pages and structure |
| Prototype Ideas | Quick idea validation, creative experiments | Start from an idea. Quest will turn it into a working prototype |
- No selection: Quest automatically determines the best approach
- Choose Code with Spec: Will always generate a Spec document
- Choose Build a Website / Prototype Ideas: Completely skips Spec for fast execution
Code with Spec Mode
Use Cases
- Feature development, bug fixes, code refactoring
- Need clear technical plans and acceptance criteria
- Team collaboration, documentation required
Workflow
1. Requirement Clarification
After inputting your task, Quest may ask clarifying questions (in multiple-choice format):- Recommend: Let Quest automatically select default answers
- Continue: Manually select and continue
- Skip: Skip questions and proceed with the conversation
2. Generate Spec
After choosing the Code with Spec scenario, Quest generates a structured Spec document:- Requirement description
- Design plan
- Task breakdown
- Acceptance criteria
3. Review Spec
- View the complete document in the Spec Tab on the right
- Modify Spec through conversation (can adjust anytime before clicking Run)
- Click Run Spec when satisfied to begin execution
4. Execute & Monitor
- To-do List: Real-time task progress (in conversation area)
- Changed Files: View code changes (in output area)
- Add requirements mid-task: Send new requirements in the input box anytime—Quest will adjust the plan
5. Review Results
After execution completes:- Accept (Local mode): Apply all changes to the workspace
- Reject (Local mode): Discard all changes
- Apply (Parallel mode): Merge to main branch
- Create PR (Remote mode): Create a Pull Request
Build a Website / Prototype Ideas Mode
Use Cases
- Quickly create website or app prototypes
- Validate product ideas
- No need for complex technical documentation
Workflow
1. Describe Your Idea
Directly describe the website/prototype you want:2. Automatic Execution
- Quest automatically selects tech stack (common web frameworks)
- No Spec needed, writes code directly
- Automatically runs preview after code completion
3. Live Preview
- Output area automatically opens Preview Tab
- View running results in real-time
- If page errors occur, Quest automatically checks and fixes
4. Iterative Refinement
Continue the conversation to adjust:Execution Environments
Quest supports three execution environments—choose based on task characteristics:Local
- Features: Modifies the main workspace directly, with zero startup cost
- Best for: Simple tasks, quick validation
- Multi-task support: Supports multiple parallel Local tasks
Parallel
- Features: Creates a hidden workspace in the background, keeps the main branch clean
- Best for: Medium complexity tasks, multi-task parallel execution
- Advantages: Unlimited Apply iterations
Remote
- Features: Remote container execution, local machine can shut down
- Best for: Complex long-running tasks, resource-intensive jobs
- Dependency: Requires a configured Git repository (currently supports GitHub)
- Multi-task support: Unlimited parallel execution
- With Git repository → All three environments supported
- Without a Git repository → Local only
Interface Layout
Three-Column Layout
Task List
- New Quest: Create a new task
- Task Status:
- Running: Executing
- Action Required: Waiting for user action
- Ready: Completed, can continue conversation
- Error: Execution error
- Tasks sorted by last interaction time
Conversation Area
- Displays conversation history
- To-do List updates in real-time (during execution)
- The input box’s bottom can compress context to optimize prompts
Output Area
- Spec Tab: View/download Spec document (Code with Spec mode)
- Changed Files Tab: View code changes, Accept/Reject
- Preview Tab: Real-time preview of running results (Website/Prototype mode)
Pricing Information
Usage Restrictions
All users need Credits to use Quest:- Pro users: Normal usage, but cannot continue when Credits run out
- Free users: Need Credits to use Quest
When Credits Run Out
- Cannot start new tasks
- If Credits run out during task execution:
- Task will terminate
- Upgrade prompt will be displayed
Common Operations
Pause/Resume Task
- While the task is executing, click the Pause button
- After pausing, click Resume to continue execution
Add Requirements
During execution, send new requirements directly in the input box:Delete Task
- Right-click the task or hover to show the menu
- Click Delete and confirm
Best Practices
Write Good Task Descriptions
Be clear and specific:- Select relevant files or code snippets
- Specify the expected tech stack
- Mention design style or references
Choose the Right Scenario
- Code with Spec: Complex features, need documentation → strict execution
- Build a Website: Quick setup → what you see is what you get
- Prototype Ideas: Validate ideas → rapid iteration
- Unsure: Don’t select a scenario, let Quest decide automatically
Choose the Right Environment
- Local: Fix a bug, adjust styling → quick feedback
- Parallel: Develop new features, refactor modules → isolated safety
- Remote: Complex refactoring, data migration → long-running stability
Use Spec Wisely
- Choose Code with Spec scenario to ensure complex tasks have clear plans
- Adjust Spec through conversation after generation, not manual editing
- For simple tasks, choose other scenarios to skip Spec
Iterative Refinement
- Let Quest complete the basic version first
- Review results, then continue the conversation to adjust
- Supports unlimited iterations until satisfied