QoderWake CN lets you build a team of digital employees (Wakers) on your local machine. Each Waker has a defined role, name, persona, and skillset. You can chat with them, assign tasks, or have them work automatically on a schedule or in response to events. Their capabilities, tools, memory, and permissions are fully configurable.
Set up QoderWake CN from scratch: install the tool, open the console, create your first Waker, assign tasks, and manage your team.
Product philosophy: 24/7 digital employees, "Always awake, always working."
Product features: Secure and controllable, production-ready, and continuously evolving.
Core concepts: Role → Agent → Agent Teams.
Initial Roles: Backend Engineer, Frontend Engineer, Test Engineer, Product Manager, Data Analyst, and Content Operations, with support for creating a Custom Role.
The invitation-only beta includes the following roles. Choose the one that best fits your needs.
QoderWake CN supports local and cloud deployment. For optimal 24/7 operation, deploy on a device that is always on and has a stable network connection, such as a dedicated local machine or a cloud desktop.
Deploy on a stable, long-running local device such as a Mac mini for reliable 24/7 operation.
QoderWake CN supports major cloud desktop and cloud server environments. For convenient browser-based login and local web management, use Elastic Desktop Service (Enterprise or Personal Edition).
QoderWake CN includes several pre-configured digital employee roles for software development. Before installation, ensure your Git environment is set up and your user account has access to the target code repository:
Go to the QoderWake CN official website to download the installation package.
Alternatively, open Terminal and run the following command:
The script automatically detects your platform, downloads and verifies the installation package, installs the main program and resource files, adds the command entry point to your
Open a terminal and run the following command:
The installation script automatically:
Default Web Console address:
Open your browser and go to
Click the "Create Waker" button on the left.
After creating a Waker, you can start a conversational task or create an automated task.
Type your request in the input box and send it. The Waker begins working immediately. During this process, you can:
Automated tasks are ideal for scenarios like "run once a day" or "automatically respond to changes in a repository."
Go to Waker Details → Triggered Tasks → New to configure the following:
Go to a Waker's details page to view and adjust its configurations.
Each Waker maintains long-term memory about you and its projects. Memory comes from three sources:
Skills are capability packages that a Waker can invoke during conversational or automated tasks.
Connectors link Wakers to external tools and services. You can add Connectors manually.
A project defines a Waker's workspace. Go to Waker Details → Projects to:
Permissions control what a Waker can and cannot do. Go to Waker Details → Permissions to configure protection categories and policies for built-in tools.
Tool Protection: Validates tool parameters based on built-in security rules. If a high-risk rule is triggered, it will request user authorization. The rule categories are:
You can approach this in two ways.
Method A: Start a conversation manually: Open the console and select a Waker with the Backend Engineer role. In the chat box, type: "Help me find the root cause of this error and provide a minimal fix," and attach the error message. The Waker will analyze the problem, identify the root cause, and propose a solution. Once you confirm, it applies the fix and submits a PR. It will ask for your approval before running commands or modifying files.
Method B: Set up an automated task: Go to Waker Details → Triggered Tasks → New. For the trigger method, select "Event" and configure it to listen for issue changes in a GitHub repository. For the task description, write: "When a new bug issue is created, automatically analyze it and create a PR with a fix." After saving, the Waker automatically starts its analysis and repair process whenever a new bug issue appears in the repository.
To report issues or suggest features, click the Feedback button in the bottom-right corner of the console. We will respond to your feedback promptly.
Which role do you need?
The invitation-only beta includes the following roles. Choose the one that best fits your needs.
| Role | Description |
|---|---|
| Backend Engineer | Specializes in API development, data modeling, service integration, performance optimization, and online stability. Follows engineering practices such as incremental delivery, test validation, and convention-first design. The Agent automatically analyzes bugs or issues, fixes the code, and submits a PR. |
| Frontend Engineer | Focuses on front-end UI design and implementation. Excels at component architecture, refining visual language, responsive design, accessibility optimization, and performance tuning. Adheres to incremental delivery, layered validation, and evidence-based verification to balance user experience with engineering quality. |
| Test Engineer | Provides quality assurance for general command-line tools and web products. Focuses on creating test plan documents, performing end-to-end testing, reproducing defects, and generating evidence-based reports. Does not run unit tests or fix problems. |
| Product Manager | An AI-native product management role for software products. Focuses on goal-driven, full-lifecycle requirements management, PRD generation, user feedback analysis, competitive research, and release communication. Sets clear approval gates for external write operations. |
| Data Analyst | An AI-native data analyst focused on problem framing, metric definitions, MCP-based data collection from DingTalk documents and spreadsheets, data diagnostics, market context analysis, and evidence-based recommendations. |
| Content Operations | An AI-native content operations role focused on account positioning, trend insights, content calendar management, creating Xiaohongshu posts, briefing visual assets, publishing after approval, engaging with comments, performance review, ensuring brand compliance, and adapting content for multiple platforms. |
| Custom | Define the role's workflow and work style yourself. You can attach any configuration, such as MCPs and Skills. |
Environment preparation
QoderWake CN supports local and cloud deployment. For optimal 24/7 operation, deploy on a device that is always on and has a stable network connection, such as a dedicated local machine or a cloud desktop.
The Windows version is not yet publicly available but is coming soon. Please use QoderWake CN on macOS or Linux for now.
Local deployment
Deploy on a stable, long-running local device such as a Mac mini for reliable 24/7 operation.
- macOS: 13.0 or later.
- Linux: Ubuntu 22.04 LTS or later is recommended. A GUI desktop environment is required to use the local Web Console.
Cloud deployment
QoderWake CN supports major cloud desktop and cloud server environments. For convenient browser-based login and local web management, use Elastic Desktop Service (Enterprise or Personal Edition).
- Linux: Ubuntu 22.04 or later. A GUI desktop environment is required to use the local Web Console.
Prerequisites
QoderWake CN includes several pre-configured digital employee roles for software development. Before installation, ensure your Git environment is set up and your user account has access to the target code repository:
gitis installed.- Repository authentication is configured.
- You can successfully run
clone,pull, andpushoperations.
Step 1: Install QoderWake CN
macOS
Go to the QoderWake CN official website to download the installation package.
Alternatively, open Terminal and run the following command:
PATH, launches a browser for login, starts the local background service, and opens the Web Console. The command entry point is: ~/.qoderwake-cn/bin/qoderwake.
Linux
Open a terminal and run the following command:
- Detects the Linux CPU architecture.
- Downloads and verifies the QoderWake CN installation package for your platform.
- Installs the QoderWake CN main program and resource files.
- Creates a command entry point and attempts to add it to your
PATH. - Launches a browser for login.
- Starts or restarts the local background service after you log in.
- Opens the local Web Console.
Windows
A Windows version is coming soon. This page will be updated upon its official release.
Step 2: Open the console
Open your browser and go to http://127.0.0.1:19830/. This is the local QoderWake CN console where you manage Wakers, assign tasks, and configure permissions.
Interface Overview:
- Left Sidebar: Displays the "My Wakers" list, showing all created Wakers and their last activity time.
- Top Buttons: Click "Create Waker" to add a new Waker. At the top right, you can create a new task.
- Main Area: After you select a Waker, the chat interface opens by default, allowing you to assign tasks directly.
- Right Sidebar: View details of created tasks and currently running tasks.
Step 3: Create a Waker
Click the "Create Waker" button on the left.
- Select a role template: Each role template comes with a pre-defined persona, a typical workflow, a core Skill set, and default prompts. Click "View Details" to preview them.
- Or create a custom role: If none of the pre-configured roles fit your needs, select "Custom Role". You can define the Waker's persona, expertise, and work style from scratch, then configure its Skills, Connectors, and permissions.
- Fill in the basic information: Enter the Waker's name, avatar, and bio. If you select a pre-configured role, this information is filled in automatically, but you can edit it as needed.
- Save: Click Save to create the Waker. After you create a Waker, an onboarding guide appears, suggesting which Skills to add and channels to connect. You can skip this and configure it later.
For best results with a Custom Role, provide comprehensive information during creation. This includes the role's positioning, core responsibilities, workflow, required Skills, and tool capabilities such as MCP configuration.
Step 4: Start working with your Waker
After creating a Waker, you can start a conversational task or create an automated task.
Conversational task
Type your request in the input box and send it. The Waker begins working immediately. During this process, you can:
- Ask follow-up questions or adjust its direction: Intervene at any time, and the Waker will continue based on the context.
- Attach files or screenshots: You can add attachments at any point in the conversation.
- Interrupt: Stop the Waker if it goes off track.
- Approve operations: When the Waker needs to modify a file, run a command, or access the internet, it pauses and requests your approval. Click Allow or Deny.
- Answer the Waker's questions: When it needs to make a choice, it will prompt you for input. Make your selection directly on the card presented.
- Switch AI model: If the current model performs poorly or you want to reduce costs, you can switch to a different model from the top dropdown menu and continue the conversation.
- View the thought process: See the Waker's reasoning in real time before it makes a decision.
- View artifacts: The right-hand panel shows the output generated by the Waker in real time, such as code patches and new files.
- Select the working directory: At the bottom of the chat, you can choose the local directory where the Waker will operate.
Automated task
Automated tasks are ideal for scenarios like "run once a day" or "automatically respond to changes in a repository."
Go to Waker Details → Triggered Tasks → New to configure the following:
- Task Name and Description: Helps the Waker understand what to do.
- AI model: Choose which model to use.
- Working directory: The local directory, repository, or project workspace where the task will be executed.
- Trigger method (up to 5, can be combined):
- Scheduled: Run once, periodically, or on a complex schedule (for example, 9 AM daily, every Monday, or the 1st of each month).
- Event-driven: Listen for changes to an issue, PR, or comment in a GitHub repository.
- Maximum Runs / End Date (optional): Prevent the task from running indefinitely.
Step 5: Manage your Wakers
Go to a Waker's details page to view and adjust its configurations.
| Section | Description |
|---|---|
| Home | View the Waker's avatar, status, days since creation, number of triggered tasks, number of conversational tasks, and activity heatmap. |
| Projects | Bind the code repositories or local directories where the Waker will work. A project can contain multiple sources (directories or repositories). Project memory is separate from the Waker's personal memory. |
| Triggered tasks | Add, delete, modify, or pause automated tasks for this Waker. |
| Conversational tasks | Review past conversations and continue chatting. |
| Memory | View and edit the Waker's long-term knowledge about you and its projects. Memory is sourced from automatic capture during conversations, manual edits, and periodic system review and deduplication. All memory is stored locally and is not uploaded to the cloud. |
| Skills | Install new Skills from the official Qoder CN Skills Marketplace, upload custom Skills from your local machine, or enable or disable existing Skills. Built-in Skills cannot be uninstalled. |
| Connectors | Integrate with external tools and services like GitHub, Jira, and GitLab. |
| Permissions | Tool Protection: Validates tool parameters based on built-in security rules to prevent threats like command injection, resource abuse, code execution, network abuse, sensitive file access, and privilege escalation. File Protection: Controls read and write access to local files. Built-in tools: Set an Allow / Ask / Deny policy for each built-in tool. For high-risk operations, an approval card is displayed. |
Waker configuration
Memory management
Each Waker maintains long-term memory about you and its projects. Memory comes from three sources:
- Automatic capture: The Waker automatically saves important information from conversations.
- Manual entry: You can go to the Memory page and edit entries directly.
- System review: The system periodically organizes and deduplicates memory entries.
Skills
Skills are capability packages that a Waker can invoke during conversational or automated tasks.
- Skills Marketplace: Install from the official Qoder CN Skills Marketplace.
- Upload Skill: Import a custom Skill from your local machine.
- Built-in Skills: Included with the system and cannot be uninstalled.
Connectors
Connectors link Wakers to external tools and services. You can add Connectors manually.
Projects
A project defines a Waker's workspace. Go to Waker Details → Projects to:
- Bind local directories or Git repositories.
- A project can contain multiple sources (directories or repositories).
- A project has its own separate memory, distinct from the Waker's personal memory.
Permission management
Permissions control what a Waker can and cannot do. Go to Waker Details → Permissions to configure protection categories and policies for built-in tools.
Tool Protection: Validates tool parameters based on built-in security rules. If a high-risk rule is triggered, it will request user authorization. The rule categories are:
- Command injection: Detects dangerous operations like
rmandmv. - Resource abuse: Detects fork bombs, system reboots, and similar threats.
- Code execution: Detects remote execution patterns like
curl | bash. - Network abuse: Detects reverse shells and local loopback access.
- Sensitive file access: Detects access to critical system files.
- Privilege escalation: Detects privilege escalation attempts like using
sudo.