Atlassian Code Context: GitHub vs. Bitbucket Setup

Quick answer

Atlassian Code Context supports GitHub Cloud and Bitbucket Cloud, but the current setup is not identical. GitHub uses the GitHub for Atlassian/Jira GitHub connector and mirrors selected repository code into Atlassian infrastructure before indexing. Atlassian says Bitbucket code is indexed directly. As of August 27, 2026, GitHub users are in open beta while Bitbucket availability is rolling out.

For either SCM, organization admins opt in at the Atlassian site level, SCM permissions are checked at query time, and indexing can lag behind recent pushes.

GitHub vs. Bitbucket

Setup area	GitHub Cloud	Bitbucket Cloud
Current rollout wording	Open beta for GitHub users	Rolling out shortly on the current support page
Connection	GitHub for Atlassian / Jira GitHub connector	Bitbucket integration
Indexing path	Code is mirrored into Atlassian infrastructure before indexing	Code is indexed directly
Indexed scope	All files across all branches and tags in opted-in repositories	All files across all branches and tags in opted-in repositories
Permission check	Existing SCM repository and branch access, validated at query time	Existing SCM repository and branch access, validated at query time
Incremental freshness	Pushes and branch updates trigger re-indexing	Pushes and branch updates trigger re-indexing

Sources: Code Context architecture and limitations and site setup, reviewed August 27, 2026.

Before connecting GitHub

1. Verify the organization and site are eligible and the Code Context enable control is visible.
2. Confirm the GitHub for Atlassian connector is installed for the intended site and organization.
3. Select only repositories your organization is authorized to mirror, index, store, and embed.
4. Review every branch and tag because the documented index scope is broader than a default branch.
5. Scan for committed credentials and rotate findings before enabling.
6. Test users with full, limited, and no repository access.

GitHub repository mirroring means the review must cover more than connector installation. Record which code and metadata cross into Atlassian infrastructure, where they are stored, how permissions are refreshed, and how deletion is verified after disablement.

Before connecting Bitbucket

Start with availability: a support page can list Bitbucket support while your site has not received the rollout. Do not build a deadline around a control that is not visible in Atlassian Administration.

When eligible, validate the exact Bitbucket Cloud workspaces and repositories, branch permissions, inherited groups, contractor access, archived projects, forks, and service accounts. Direct indexing does not eliminate the storage of code chunks, metadata, and vector embeddings described by Atlassian.

Shared permission boundary

Atlassian says query results are limited to repositories and branches the user already can access in the SCM, with permission validation at query time. That makes SCM access review the foundation of rollout.

Test at least these cases:

• A user who lost repository access.
• A user who can read one branch but not another.
• A newly private or transferred repository.
• A connector outage or revoked credential.
• A recent push that has not yet reached the index.
• A repository with stale or misleading comments.

Permission-aware retrieval does not prove the answer is relevant or correct. Semantic results have no current confidence signal, and Atlassian warns that indexing is frequent rather than instant.

Initial indexing and freshness

The setup guide says initial indexing can take up to 24 hours and there is currently no indicator showing which repositories are indexed. Define a pilot acceptance test using known exact strings, semantic questions, recent commits, deleted files, branches, and tags.

Do not use “no result” as proof that code does not exist until the connector, permission endpoint, indexing window, query mode, and repository scope have been checked.

Disable and recovery test

Atlassian says disabling Code Context stops indexing immediately and deletes stored embeddings, indices, and code chunks. Before production adoption, run a controlled disable test and document:

1. Who can disable the feature and connector.
2. What users see during and after disablement.
3. How your organization obtains or records deletion assurance.
4. Whether cached third-party agent context or logs require separate deletion.
5. How access is safely restored without silently expanding repository scope.

Use the Code Context security guide and readiness check before enabling either SCM.

Frequently asked questions

Is GitHub setup available before Bitbucket?

The current support page says open beta for GitHub users and Bitbucket rolling out shortly. Your admin console is the source of truth for site eligibility.

Does Code Context index only the default branch?

No. Atlassian documents all source files across all branches and tags in opted-in repositories.

Can contributors exclude individual files?

Atlassian says individual contributors cannot exclude their code once an organization admin enables Code Context for a repository. Resolve scope and rights before opt-in.

Does SCM permission enforcement cover third-party coding agents?

It bounds retrieval from Code Context. Once source is sent to a third-party agent through the TWG CLI, that provider’s terms, retention, security, costs, and agent actions also apply.

Official sources

• Atlassian Support: what is Code Context
• Atlassian Support: set up Code Context
• Atlassian: introducing Code Context

Source check: August 27, 2026. Recheck rollout, connectors, supported SCMs, indexing scope, storage, permission behavior, and deletion terms.