Overview
EdgeOne Pages Copilot is an AI programming assistant integrated in Pages. You just need to use natural language description of requirements, and Copilot can understand code, generate modification solutions, complete code change, and help you publish, all combined with project context.
Regardless of whether you want to quickly create a webpage, fix an online issue, or continuously iterate on an existing project, Copilot can help you complete tasks more efficiently.
Core Capabilities
Language-driven development: No need to possess deep development background, just start with simple dialogue prompt content.
Understand project context: Copilot combines project code and build configuration to understand the current site, outputting modification results closer to the project's current status.
Iteration for existing projects: Particularly suitable for updating, fixing, and optimizing templates, Github, and Pages Projects.
Standalone preview environment: Each time Copilot executes a modification, an independent preview deployment is automatically generated. It will be published to the production environment only after confirmation, ensuring production security.
Convenient access on mobile terminal: Support initiating modification directives anytime, anywhere on mobile phones and other devices, making it easy to complete tasks like content updates.
Typical Scenarios
Quick Repair and Routine Maintenance
Suitable for processing tasks with clear scope and expected outcomes, such as:
Fix styling issues, layout misalignment, or interactive anomalies on the page.
Correct copywriting errors, broken links, or resource reference issues.
Content Update and Iteration
Marketing, design and content teams can directly edit page copywriting, replace images and adjust theme color through dialogue—without the need to intervene developers.
Quick Creation of a New Page
Ideal for generating a landing page or event topic page for an existing project in one sentence.
Intelligent Troubleshooting
When a build or deployment fails, you can directly initiate Copilot analysis based on failure records, allowing it to assist in identifying issues combined with error logs.
Use Limits
Currently Copilot is in free public testing phase. The number of executable tasks per day is subject to factors such as selected programming mode, task complexity, code repository scale and session context length.
Programming Mode | Model Capability | Number of Tasks / Day |
Expert mode | Underlying use of top-level large models such as OpenAI Codex | 5-10 |
Standard Mode | Underlying use of high-quality models such as GLM | 20-30 |
When the quota of Expert mode is exhausted, the system will automatically switch to Standard mode. When the quota of Standard mode is exhausted, it will automatically refresh the next day.
Using Pages Copilot
Automatic Analysis of Failed Deployment
If the project build fails, you can click
Why did it fail on the details page of the failed deployment in the console, allowing Copilot to assist in analyzing the reason based on the failure log and continue the fix.Note:
Copilot currently only supports GitHub projects.
