YouTip

QoderWake allows you to build a team of digital employees (Wakers) on your local machine. Each employee has a position, name, persona, and expertise. They can be summoned for chat or work, and can also start working automatically based on schedules or events. Core product philosophy: 7x24 digital employees, "Always awake, always working." !(#) QoderWake is a digital employee management platform that supports both local and cloud deployment. * Core features: secure and controllable, production-ready, continuously evolving. * Core concept: three-layer progression from Employee Role to Digital Employee (Agent) to Team/Organization (Agent Teams). !(#) After registration, click the Download button in the upper right corner: !(#) Select the QoderWake option and download the installer according to your computer system: !(#) After downloading, double-click to install. The installation completion screen is as follows: !(#) The settings button in the lower left corner can be used to set the language. After switching to Chinese, it looks like this: !(#) ### Available Roles Overview The current beta covers the following roles. Choose the one that best fits your needs: | Role | Position Characteristics | | --- | --- | | Backend Engineer | Focuses on API development, data modeling, service integration, performance optimization, and online stability. Follows engineering practices of incremental delivery, test validation, and convention-over-design. Automatically analyzes bugs/issues and submits PRs after fixing code. | | Frontend Engineer | Focuses on frontend interface design and implementation, excelling in component architecture, visual language refinement, responsive adaptation, accessibility optimization, and performance tuning. Follows incremental delivery, layered validation, and evidence-based verification, balancing user experience and engineering quality. | | QA Engineer | A quality assurance role for general command-line tools and web products, focusing on test plan documentation, end-to-end testing, bug reproduction, and evidence-based reporting; does not run unit tests or fix issues. | | Product Manager | An AI-native product management role for software products, focusing on goal-driven full-lifecycle requirement management, PRD generation, user feedback analysis, competitive research, and release communication, with explicit approval gates for external writes. | | Data Analyst | An AI-native data analyst for problem framing, metric definitions, data collection, data diagnosis, market context, and evidence-based recommendations. | | Content Operations | An AI-native content operations role for account positioning, trend insights, content calendar, Xiaohongshu note creation, visual asset briefs, post-approval publishing, comment engagement, data review, brand compliance, and cross-platform rewriting. | | Custom | Define the position's working methods and workflows yourself, with support for mounting MCP, Skill, and other configurations. | * * * ## Installation Environment Requirements QoderWake supports both local and cloud deployment. It is recommended to deploy on a stable, always-online device with good network connectivity or a cloud desktop environment to ensure 7x24 continuous operation of digital employees. > Windows version coming soon, not yet available. Please use macOS or Linux first. ### Local Deployment Requirements Recommended to deploy on a locally stable, long-running device such as a Mac mini for a better 7x24 operation experience. | System | Version Requirements | | --- | --- | | macOS | Version 13.0 and above | | Linux | Recommended Ubuntu 22.04 LTS and above, GUI desktop required for local Web Console | ### Cloud Deployment Requirements Supports mainstream cloud desktop and cloud host environments. For convenient browser login and local web management, Alibaba Cloud Wuying Cloud Computer (Enterprise or Personal edition) is recommended. ### Prerequisites The product comes with multiple software development digital employee roles. Before installation, please prepare the Git environment and confirm that the current account can access the target code repository: * Git installed * Repository authentication configured * Able to execute clone, pull, and push operations normally Both local and cloud deployment environments require external network access for user login, model invocation, code repository access, etc. In enterprise networks, proxy, or firewall environments, please ensure necessary HTTPS outbound access is allowed in advance. * * * ## Step 1: Install QoderWake Choose the installation method according to your operating system. ### macOS Installation Method 1: Direct download of installation package. * Apple chip (ARM): (https://download.qoder.com/qoderwake/installers/latest/qoderwake_latest_darwin_arm64.dmg) * Intel chip: (https://download.qoder.com/qoderwake/installers/latest/qoderwake_latest_darwin_amd64.dmg) Method 2: One-click installation via terminal. # One-click installation command, automatically detects platform and completes installation curl -fsSL https://qoder-ide.oss-ap-southeast-1.aliyuncs.com/qoderwake/install.sh | bash The script will automatically: detect platform, download and verify installation package, install main program and resource files, add command entry to PATH, launch browser login, start local resident service, and open Web Console. After installation, the command entry is: ~/.qoderwake/bin/qoderwake ### Linux Installation Open terminal and run the same one-click installation command as macOS. The installation script will automatically complete the following steps: 1. Detect Linux CPU architecture 2. Download and verify QoderWake installation package for the corresponding platform 3. Install QoderWake main program and resource files 4. Create command entry and attempt to add to PATH 5. Launch browser login 6. After login, start or restart local resident service 7. Automatically open local Web Console Command entry: ~/.qoderwake/bin/qoderwake Default Web Console address: http://127.0.0.1:19820 * * * ## Step 2: Open Console Access http://127.0.0.1:19820/ in your browser. This is QoderWake's management console on your local machine. Employee management, task assignment, and permission configuration are all done here. !(#) ### Interface Layout Description | Area | Function Description | | --- | --- | | Left sidebar | Displays "My Waker" list, showing all created employees and recent activity time | | Top buttons | Click "Create Waker" to quickly create a new employee; new task creation supported on the right | | Main area | After entering an employee, the conversation page opens by default for direct task assignment | | Right sidebar | View created tasks and details of currently running tasks | First visit will guide you to log in to your account. Follow the prompts to complete. * * * ## Step 3: Create Digital Employee Click the "Create Waker" button on the left to enter the creation process. !(#) ### Creation Process 1. Choose a role template: Each role template comes with persona, typical workflow, core skill set, and default questions. Click "View Details" to preview. 2. Or custom role: If preset roles don't fit, choose "Custom Role" to fully define this employee's persona, expertise, and working style, then configure skills, connectors, and permissions. 3. Fill in basic information: Enter employee name, avatar, and introduction. Content will be auto-populated when selecting a preset role and can be adjusted as needed. 4. Save: Click save to create the employee. After first creation, the new employee will enter onboarding, recommending which skills to mount and channels to connect. You can skip and configure later. > To improve custom role effectiveness, it is recommended to provide complete role information during creation: role positioning, core responsibilities, workflow, required skills, and MCP and other tool capability configurations. * * * ## Step 4: Start Working with Your Employee After new employee creation is complete, you can assign work via Direct Conversation or Create Automated Task. ### Conversation Task Simply enter requirements in the input box and send, and the employee starts working. !(#) Operations you can perform during the process: | Operation | Description | | --- | --- | | Follow-up, adjust direction | Interject anytime; employee continues with context | | Attach files/screenshots | Attachments can be added to any round of conversation | | Interrupt | Stop the employee if they go off track | | Approval operations | Employee pauses for consent when modifying files, running commands, or going online; you click "Allow" or "Deny" | | Answer employee questions | Employee will ask you when facing choices; select directly on the card | | Switch AI model | If the model seems insufficient or you want to save costs during conversation, switch models from the top dropdown | | View thinking process | Real-time view of employee's reasoning before making decisions | | View outputs | Right panel shows employee-produced code patches, newly generated files, etc. in real time | | Select working directory | Choose the employee's local working directory at the bottom of the conversation | ### Automated Task Suitable for scenarios like "run once a day" or "automatically respond when a repository changes." !(#) Go to Waker Details - Trigger Tasks - New to configure: | Configuration Item | Description | | --- | --- | | Task name and description | Employee understands what to do upon reading | | AI model | Select the model to use | | Working directory | Which local directory, repository, or project workspace to execute in | | Trigger method (up to 5) | Scheduled (one-time/periodic/complex periodic), event-driven (monitor GitHub repository changes), webhook (third-party systems trigger via fixed address) | | Max runs/deadline (optional) | Limit task from running infinitely | After saving, the employee works automatically according to your set schedule. You can pause/enable, run immediate tests, and view run history at any time. * * * ## Step 5: Manage Your Digital Employees Enter the employee details page to view and adjust all configurations. | Section | Function Description | | --- | --- | | Home | View avatar, status, days employed, trigger task count, conversation task count, work heatmap | | Projects | Bind code repositories or local directories for them to work on. One project can contain multiple sources (multiple directories or repositories); project memory is separate from employee personal memory | | Trigger Tasks | Add/delete/edit/pause automated tasks on them | | Conversation Tasks | Browse historical conversations and continue chatting | | Memory | View/edit their long-term understanding of you and projects. Sources: automatically accumulated during conversation, manually edited, system periodically organizes and deduplicates. Stored locally, not uploaded to cloud | | Skills | Install from Qoder official Skills Marketplace, upload custom skills locally, or toggle existing skills. Built-in skills cannot be uninstalled | | Connectors | Connect to external tools/services (GitHub, Jira, GitLab, etc.) | | Permissions | Tool protection, file protection, built-in tool policies. High-risk operations will trigger approval cards | * * * ## Employee Configuration Details ### Memory Management Each employee has their own long-term memory for accumulating understanding of you and projects. There are three sources of memory: | Source | Description | | --- | --- | | Automatic accumulation | Information deemed worth remembering during conversation is automatically written | | Manual entry | You can directly edit from the memory page | | System organization | Periodically organized and deduplicated by the system | Go to Employee Details - Memory to browse, search, and edit all memory entries. Memory is stored locally and not uploaded to the cloud. ### Skills Skills are specialized capability packages that employees can invoke during conversations or trigger tasks. * Skills Marketplace: Install from Qoder official Skills Marketplace * Upload Skill: Import locally * Built-in skills: System-provided, cannot be uninstalled Go to Employee Details - Skills to view installed skills, search marketplace, and upload custom skills. ### Connectors Connectors are bridges for employees to connect with external tools/services, supporting manual addition of connectors. ### Projects Projects are employee workspaces. Go to Employee Details - Projects to bind local directories or git repositories. One project can contain multiple sources (multiple directories/repositories). Projects have independent memory, separate from employee personal memory. ### Permission Management Permissions control what employees can and cannot do. Go to Employee Details - Permissions, which includes two categories of protection plus built-in tool policies. #### Tool Protection Validates tool parameters based on built-in security rules. When high-risk rules are hit, user authorization is requested. | Rule Category | Detection Content | | --- | --- | | Command injection | Detects rm, mv and other dangerous operations | | Resource abuse | Detects fork bombs, system reboots, etc. | | Code execution | Detects curl | bash and other remote executions | | Network abuse | Detects reverse shells, localhost access, etc. | | Sensitive file access | Detects access to system critical files | | Privilege escalation | Detects sudo and other privilege escalation operations | #### File Protection Controls employee read/write permissions for local files. #### Built-in Tools Set Allow/Ask/Disable policies for Waker's available built-in tools. > When employee operations trigger security rules, an approval card will pop up for you to decide whether to allow. * * * ## Scenario Example: Automatic Bug Fixing Two driving methods can achieve automatic bug fixing. ### Method A: Manually Initiate Conversation 1. Open console, select a Waker with Backend Engineer role 2. Enter requirements in the dialog box and attach error information 3. Employee analyzes problem, locates root cause, and proposes fix 4. After confirmation, have them execute the fix and submit PR ## Example # In the dialog box, enter something like the following, attaching error logs # "Help me investigate the root cause of this error and provide a minimal fix" # The employee will perform operations similar to: # 1. Analyze error logs and locate problematic code # 2. Provide fix plan for your confirmation # 3. Execute code fix # 4. Submit PR to code repository When executing commands or modifying files, they will first ask for your consent. ### Method B: Set Up Automated Task Processing 1. Go to Employee Details - Trigger Tasks - New 2. Select "Event" as trigger method, monitoring GitHub repository Issue changes 3. Task description: "When a new Bug Issue comes in, automatically analyze and submit a fix PR." 4. After saving, whenever a new Bug Issue appears in the repository, employee automatically starts analysis and fixing > Both methods have applicable scenarios: Method A is suitable for temporary, one-time problem investigation; Method B is suitable for continuous automated operations, reducing manual intervention. * * * ## Frequently Asked Questions ### Cannot open Web Console after installation? Confirm whether local service is running normally, and check if http://127.0.0.1:19820 is accessible. If there is a port conflict, check if another program is using port 19820. ### Employee freezes during task execution? Check if employee needs your approval confirmation, and see if there are pending approval cards in the conversation window. ### Custom role not performing well? It is recommended to provide complete role information during creation: including role positioning, core responsibilities, workflow, required skills, and MCP and other tool capability configurations. The more complete the information, the better the results. ### How to provide feedback? For any issues or product feature feedback, click the feedback button in the lower right corner of the console to submit. The team will respond and resolve as soon as possible.