GitHub Copilot in the Browser: A Practical Workflow

You can reach GitHub Copilot in a browser without installing any software. The two main entry points are the Copilot Chat panel on GitHub.com and GitHub Codespaces, which gives you a full VS Code environment in a browser tab. Tools like fenbs let you carry the approved task brief between browser sessions and local IDE sessions, so whichever environment you pick up the work in starts from the same context.

Entry points for browser-based Copilot

GitHub.com hosts a Copilot Chat interface accessible from the navigation bar at the top of the page. You can ask questions about a repository, a specific file, or a recent pull request without leaving the browser. The chat interface knows about the repository you are viewing and can reference code from it in its answers.

GitHub Codespaces opens a VS Code-compatible editor in the browser, provisioned in the cloud. It has full access to your repository and runs the Copilot extension just as it would in a local IDE. You get inline completions, chat, and the ability to run code. Codespaces is useful when you are working on a machine that does not have your development environment installed, or when you need a clean, reproducible environment for a specific task.

Copilot Chat on GitHub.com

The Copilot Chat panel on GitHub.com can answer questions about the repository you are currently looking at: what a function does, how a feature is implemented, or what changed in a recent commit. You can ask it to summarize an issue or explain a pull request diff. It does not have write access to the repository; it can only read and explain.

The context available to the GitHub.com chat interface is limited to what GitHub can index from the repository. For detailed questions that require understanding many files at once, a Codespaces environment or a local IDE with the Copilot extension usually gives better results because more context is available.

GitHub Codespaces and browser editing

Codespaces provisions a development container in the cloud and opens it as a VS Code instance in the browser. Your repository is cloned into the container and you can edit files, run tests, install dependencies, and commit and push changes, all from a browser tab. Copilot works in this environment exactly as it does in a locally installed VS Code.

Codespaces environments are not always free; usage is metered by compute hours and storage. Check your plan's included hours on GitHub.com before starting a Codespace for a long task. For lighter work, the built-in GitHub web editor (press the period key on any repository page) opens a VS Code interface without compute, but Copilot features are more limited in that environment.

Code review with Copilot in the browser

On a pull request page, Copilot can summarize the changes and help you understand what a diff does without reading every line. Open the Files Changed tab and use the Copilot Chat panel to ask questions about specific files or the overall purpose of the PR. This is faster for a reviewer who needs to understand a large change quickly.

Copilot cannot approve or merge pull requests, add review comments automatically, or see CI check results. The review is still yours to complete. For AI-assisted PR automation, see GitHub Copilot hooks.

Limitations of the browser compared to a local IDE

The browser-based Copilot experience is good for reviewing code, asking questions, and making focused edits. It is less practical for long sessions that require running a local build server, accessing system resources, or working with large numbers of files simultaneously. For anything that needs a full development environment running locally, a locally installed IDE with the Copilot plugin remains the better choice.

Browser sessions also end: if your Codespace times out or you close the tab without committing, any unsaved work is lost. Commit frequently and update your task record before ending a browser session.

How to do this with fenbs

1. Create a free account and add a project for the repository you are working on.
2. Write the task brief as a fenbs task before opening a browser Copilot session. Include the goal, the acceptance criteria, and any constraints.
3. At the end of a browser session, update the task with what was completed, what was committed, and what remains.
4. When you continue the work in a different environment, read the task status in fenbs before starting so you pick up from the current state rather than re-reading a long git log.
5. Connect a supported AI assistant from Integrations and ask it to read the project context, including rules and Where we left off, at the start of each session.

fenbs does not connect to GitHub Codespaces or the GitHub.com Copilot interface directly. It keeps the task and project context in a portable record. For Copilot CLI use, see GitHub Copilot CLI. For alternative AI coding tools, see GitHub Copilot alternatives.

If your team uses both browser-based Codespaces and local IDE work on the same project, a shared project record in fenbs is especially useful: it removes the need to reconstruct context each time a developer switches environments or picks up where another left off.