SI Agent News
DevelopingResearch

GitHub Copilot app adds support for Enterprise Server accounts

The update allows users to connect on-premises code repositories to the Copilot app, expanding how teams access private data.

1.2.0Minimum Copilot app version required

The short version

Why it matters

This update bridges the gap between private, on-premises code storage and mobile developer tools. It allows organizations to maintain strict repository control while still giving engineers access to their work on the go.

Moving out of preview

GitHub officially transitioned GitHub Enterprise Server (GHES) account support to general availability on November 13, 2026. This change marks the end of the opt-in preview period for the GitHub Copilot app. Users can now integrate their server-hosted environments into their daily workflow more seamlessly.

The update specifically applies to the Copilot app version 1.2.0 or later. By making this feature standard, GitHub is catering to organizations that prefer to host their own code rather than relying solely on public cloud infrastructure. This move helps align the app with the needs of large-scale corporate and academic users.

As CASRAI reported, this expansion allows the app to interact with various components of a server-hosted setup. This includes not just the code itself, but also the associated issues and pull requests. It provides a more unified experience for developers who manage complex projects across different hosting environments.

How to connect accounts

Setting up the connection requires a specific sequence of steps within the app settings. Users must first sign in with a standard GitHub.com or GHE.com account. Once that is established, they can navigate to the settings menu to add their specific Enterprise Server account.

The process involves navigating to the Settings section, selecting Accounts, and then choosing the option to Add account. From there, the user must complete a browser authorization step to link the two entities. This ensures that the app has the necessary permissions to pull data from the private server.

It is important to note that while account connection is now standard, cloud sessions are still unavailable for GHES accounts. This means that while the app can see the data, the way it interacts with that data is governed by specific server rules. The distinction between account access and session type is a key technical detail for administrators.

Models and data processing

A common point of confusion involves where the actual AI model processing takes place. Connecting a GHES account does not automatically mean that the AI models are running on-premises. The account connection is a matter of data access, while the model provider is a separate configuration choice.

GitHub offers a public preview for users who want to supply their own model providers. This allows teams to use models hosted either locally or in the cloud. When using an external provider, a GitHub account is still required, but a specific Copilot plan might not be necessary for that specific configuration.

Because these features are separate, organizations must evaluate their setup in pieces. They need to decide which repository account to connect, which model provider will handle the requests, and whether that specific combination meets their security standards. The general availability of the account connection does not establish a blanket approval for data processing locations.

Impact on research teams

For research software engineers who work with institution-hosted repositories, this update provides immediate utility. They can now interact with their research code, issues, and pull requests through the Copilot app. This facilitates a more mobile and flexible way to manage research software development.

Research offices must be careful to distinguish between these different technical layers. They need to understand that connecting an account is not the same as establishing on-premises inference. Each part of the stack—the account, the model, and the data flow—requires its own assessment for institutional requirements.

Data governance teams will need to look beyond the simple fact that the app now supports GHES. They must verify the specific configuration chosen by the researchers. This includes identifying where the data is processed and ensuring it aligns with the institution's privacy and security policies.

Administrative oversight

While the app provides a way to access the data, the heavy lifting of administration remains on the web platforms. License assignments and administrative features are still managed through GitHub.com or GHE.com. This includes the ability to set policies and manage various settings for the organization.

This means that server-hosted repositories do not represent the entire Copilot service arrangement. The app acts as a window into the data, but the rules governing that data are still dictated by the central server management tools. Administrators can still control who has access and what permissions are granted from the main dashboard.

Moving forward, organizations will need to balance the convenience of the app with the security of their server-hosted environment. The goal is to provide developers with the tools they need without compromising the integrity of the private codebases. This release is a step toward that balance by making the connection more accessible to everyone.

What to watch

Try it

Check your Copilot app version to ensure it is 1.2.0 or later before attempting to add a GHES account.

Sources and links

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

Listen to the AI Wire headlines

Listen to the AI Wire

Listen to our daily AI shows every morning: on your commute, in the shower, wherever you are

Listen to the daily shows

First reported by CASRAI · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

More AI news