Advanced Analytica Advanced Analytica: IBOM
← All connectors
GitHub logo

Source connector

GitHub

Set up GitHub as a source and move its data into your chosen destination.

Version 2.8.6 certified generally available

GitHub

Use this page to set up the GitHub source connector and find reference information.

Prerequisites

  • A list of GitHub repositories, with access to any private repositories

For Airbyte Cloud:

For Airbyte Open Source:

Setup

Step 1: Set up GitHub

Create a GitHub Account.

Airbyte Open Source additional setup steps

Sign in to GitHub, then generate a personal access token. To distribute API quota usage across tokens, enter multiple tokens separated by ,.

Step 2: Set up the connector

Airbyte Cloud

  1. Log into your Airbyte Cloud account.

  2. Select Sources, then select + New source.

  3. On the Set up the source page, select GitHub from the Source type dropdown.

  4. Enter a name for the GitHub connector.

  5. Choose an authentication method:

    • For Airbyte Cloud: Authenticate your GitHub account to authorise your GitHub account. Airbyte authenticates the GitHub account you’re currently signed in to. Check that you’re signed in to the correct account.
    • For Airbyte Open Source: Authenticate with Personal Access Token. To generate a personal access token, sign in to GitHub and generate a personal access token. Enter your GitHub personal access token. To distribute API quota usage across tokens, enter multiple tokens separated by ,.
  6. GitHub Repositories - Enter a list of GitHub organizations or repositories. For one repository, use airbytehq/airbyte. For multiple repositories, use airbytehq/airbyte airbytehq/another-repo. To receive data from every repository in an organization, use airbytehq/*.

    :::caution The connector skips repositories that don’t exist or have an incorrect name or name format. It records an WARN message in the logs. :::

  7. Start date (Optional) - Enter the date from which you want to replicate stream data. Supported streams replicate only data generated on or after this date.

    • These streams sync only records generated on or after the Start Date: comments, commit_comment_reactions, commit_comments, commits, deployments, events, issue_comment_reactions, issue_events, issue_milestones, issue_reactions, issues, project_cards, project_columns, projects, pull_request_comment_reactions, pull_requests, pull_request_stats, releases, review_comments, reviews, stargazers, workflow_run_attempts, workflow_runs, workflows.

    • The Start Date doesn’t apply to these streams, which sync all data: assignees, branches, collaborators, issue_labels, organizations, pull_request_commits, repositories, tags, teams, users

  8. Branch (Optional) - Enter the GitHub repository branches from which you want to pull commits, such as airbytehq/airbyte/master. If you don’t specify branches for a repository, the connector pulls its default branch, such as airbytehq/airbyte/master airbytehq/airbyte/my-branch.

  9. API URL (Optional) - For a self-hosted GitHub instance, enter its API URL, such as https://github.company.org. Leave this empty to use https://api.github.com/.

  10. Max Waiting Time (in minutes) (Optional) - Set how long the connector waits when every configured API token is rate-limited. After this period, the sync fails. The default is 120 minutes. This covers GitHub’s 60-minute rate-limit reset window and adds a margin. You can enter any value from 1 to 240 minutes. With multiple personal access tokens, the connector rotates through them before waiting. It waits only after exhausting every token.

  11. Number of Concurrent Threads (Optional) - Set how many partitions the connector reads in parallel. The default is 4, and the maximum is 25. During the declarative migration, this setting applies only to streams already moved to the declarative manifest. These are repositories, assignees, branches, collaborators, issue_labels, tags, organizations, teams, users, events, pull_requests, commit_comments, issue_milestones, stargazers, projects, issue_events, deployments, workflows, comments, issues, review_comments, pull_request_commits, project_columns, project_cards, team_members, team_memberships, issue_timeline_events, commit_comment_reactions, issue_comment_reactions, commits, contributor_activity, workflow_runs and workflow_jobs. The connector still reads the remaining streams one at a time. A higher value speeds up migrated streams but increases the risk of reaching GitHub’s secondary rate limits.

Airbyte Open Source

  1. Open the Airbyte Open Source dashboard.
  2. Select Sources, then select + New source.
  3. On the Set up the source page, select GitHub from the Source type dropdown.
  4. Enter a name for the GitHub connector.

Supported sync modes

The GitHub source connector supports these sync modes:

Supported streams

The connector provides these full refresh streams:

The connector provides these incremental streams:

Entity-relationship diagram

Notes

  1. Only four streams listed above are fully incremental: comments, commits, issues and review comments. They:

    • read only new records;
    • output only new records.
  2. The workflow_runs, workflow_jobs and workflow_run_attempts streams are almost fully incremental. They:

    • read new records and some old records from the past 32 days docs. GitHub prevents reruns 30 days after a run’s creation. The connector adds a two-day margin because older runs can no longer change;
    • use the same logic because workflow_jobs depends on workflow_runs to read data docs;
    • output only new records.

    workflow_runs contains only the latest attempt for each run. The GitHub endpoint returns one record per run. When a workflow reruns, earlier attempts never reach your destination. The returned record combines the latest attempt’s status with the original attempt’s created_at. Sync workflow_run_attempts when you need the complete rerun history. It emits one record per attempt, keyed by [id, run_attempt]. Each attempt has its own created_at, conclusion, logs_url and jobs_url.

    GitHub has no endpoint for listing attempts, so each rerun attempt costs one request. A run that was never rerun adds no request beyond the run listing. Its only attempt comes from the listing record, with logs_url and jobs_url scoped to that attempt. Its created_at and updated_at then match workflow_runs. These can be one second earlier than the same attempt’s values from GitHub’s attempt endpoint. Its pull_requests reflects the listing read. For an active run, its status also reflects that read, as in workflow_runs. New runs created during the listing read can cause a run to appear twice, as in workflow_runs. Destinations that deduplicate on the primary key aren’t affected. The connector selects runs for the window using the same approach as workflow_runs. It rereads a run when that run’s updated_at changes. It then emits every attempt for that run, including attempts last updated before the cursor.

    Plan for these three points before enabling it:

    • The first sync is the expensive one, and a repository is all-or-nothing. It reads every run created since the Start Date. Allow one listing request per 100 runs and one request for each rerun attempt. An authenticated token allows 5,000 requests per hour. This work doesn’t carry across syncs. A repository’s cursor advances only after the connector reads that repository to the end. If interrupted part-way through, the next sync restarts that repository. It doesn’t reread repositories it completed. Enabling the stream on a connection already syncing workflow_runs still starts from the Start Date. The stream keeps its own cursor.
    • The run listing is read once per selected stream. workflow_runs, workflow_jobs and workflow_run_attempts each read back to 32 days before their own cursor. The streams can’t share a read yet.
    • More tokens help only if they belong to different accounts. GitHub’s 5,000-request hourly limit applies per account. Its secondary rate limits also apply per account, so extra tokens for one account add no capacity. Beyond that limit, the connector applies one global budget of 900 requests per minute. This caps throughput at about eleven tokens’ capacity.

    If the token can’t read Actions for a repository, GitHub returns 403. The connector logs the issue and skips that repository without failing the sync. An empty workflow_run_attempts can therefore indicate a missing workflow scope for a classic token. It can also indicate a missing Actions: read permission for a fine-grained token, rather than no reruns.

  3. The other 19 incremental streams differ in one way. They:

    • read all records;
    • output only new records.

    Consider this behaviour when using these streams because it may affect your API request limits.

  4. For large streams, a start_date far in the past can cause repeated GitHub errors instead of records. The logs show the relevant WARN message. A more recent start_date may help. The “Start date” configuration option does not apply to the streams below, because the GitHub API does not include dates which can be used for filtering:

    • assignees
    • branches
    • collaborators
    • issue_labels
    • organizations
    • pull_request_commits
    • tags
    • teams
    • users
  5. Adding a repository or organization to a connection that has already synced doesn’t backfill its history. See Adding repositories or organizations to an existing connection.

IP allow list

If you use Airbyte Cloud and restrict access to specific IPs, add the Airbyte Cloud IP addresses to your allow list.

Limitations and troubleshooting

Expand to see details about GitHub connector limitations and troubleshooting.

Connector limitations

Rate limits

You can make API requests with a personal access token. You can also authorise a GitHub App or OAuth app to make API requests for you. These requests count towards your personal limit of 5,000 requests per hour. The limit is 15,000 for apps owned by a GitHub Enterprise Cloud organization.

:::info GitHub counts REST API and GraphQL API rate limits separately. The REST API uses a request-based limit. The GraphQL API uses a point-based limit, where each query costs a calculated number of points. Streams using the GraphQL API include pull_request_stats, reviews, pull_request_comment_reactions, issue_reactions, releases and projects_v2. :::

:::tip If you reach the limits before reading every stream, try these steps:

  1. Use Incremental sync mode.
  2. Increase the sync interval.
  3. Split the sync into separate connections with fewer streams.
  4. Enter multiple personal access tokens in the Personal Access Tokens field, separated by commas. The connector rotates through every token before waiting for rate limits to reset.

:::

When every configured token is rate-limited, the connector waits for the limits to reset instead of failing immediately. The Max Waiting Time (in minutes) option caps this wait. The default is 120 minutes, and the maximum is 240 minutes. The cap applies when the connector already knows its tokens are exhausted. It also applies when GitHub rejects a request and specifies a wait. Airbyte treats rate-limit exhaustion as a transient error. If the connector exceeds this wait, Airbyte retries according to your connection’s retry behaviour. With several tokens, a rate-limited request switches tokens instead of waiting, regardless of this setting.

Test connection is exempt from the wait. It returns the rate-limit message within seconds instead of waiting. A rate-limited token therefore doesn’t leave the connection setup page hanging.

See the GitHub article Rate limits for the REST API.

Add to an existing connection

Expanding the GitHub Repositories field after a connection has synced doesn’t backfill new repositories’ or organizations’ earlier activity. The connector keeps one sync position for each repository or organization. A new one starts from the connection’s current overall position, not your configured Start date. It never emits items created or last updated before that position. The sync provides no warning or error.

For example, consider a connection syncing docker/* since 2026-01-01. If you expand it to docker/*, airbytehq/*, it collects airbytehq repositories updated after 2026-01-01. It doesn’t collect repositories whose latest update is older.

To collect a newly added repository’s or organization’s full history, save the new value first. Then clear the affected streams, or refresh the connection, and sync. Each stream rereads from the start of its supported range. Streams respecting Start date use your configured value. The streams listed above that ignore it read everything available.

This currently affects the repositories and workflow_run_attempts streams. Other streams still use the Start date for a repository they haven’t seen. They’ll adopt the behaviour above when they move to the connector’s declarative implementation.

GitHub Enterprise Server without rate limits

GitHub Enterprise Server disables HTTP API rate limits by default. In this state, the instance returns 404 Rate limiting is not enabled. for GET /rate_limit. The connector interprets this as an instance that doesn’t track quotas. It continues without quota tracking and no longer treats the response as a failed connection. Requests still require authentication, and the connector still rotates through multiple tokens. It also respects any enforced rate limits through its usual retry and backoff. Secondary rate limits use a separate GHES setting.

Releases asset limit

The Releases stream uses the GitHub GraphQL API and retrieves up to 100 assets per release. A release with more than 100 assets includes only the first 100. The connector doesn’t currently support sub-pagination for release assets.

Stargazers access requirements

Since July 2026, GitHub only lists a repository’s stargazers for its admins and collaborators (GitHub changelog). GitHub returns 404 or 403 when the authenticated user or app has neither role. The connector skips only that repository’s stargazers stream. It logs the repository name and restriction, while the sync still succeeds. Other streams for that repository aren’t affected. Expect an empty stargazers stream for public or third-party repositories where you aren’t a collaborator. The aggregate star count remains available in the stargazers_count field of the repositories stream.

Unreadable repositories and server errors

If a repository-specific issue prevents access, the connector logs it and continues to the next repository. One unreadable repository therefore doesn’t fail the entire stream. This includes a deleted or renamed repository (404), or one your token can’t access (403). It also includes a repository without commits (409), or one where the stream’s supporting feature is disabled. For example, this applies to the issue_labels stream when Issues is disabled (410).

Version 2.2.0 changed how the connector handles repeated server errors. If GitHub continues returning 502 Bad Gateway or 504 Gateway Timeout after retries, the stream now fails. It no longer skips the affected repository or organization. Earlier versions logged a warning, skipped it and reported a successful sync. As a result, records could be missing without notice. Affected streams are repositories since 2.2.0; assignees, branches, collaborators, issue_labels, tags, organizations, teams and users since 2.3.0; events, pull_requests, commit_comments, issue_milestones, stargazers, projects, issue_events, deployments, workflows, comments, issues and review_comments since 2.4.0; pull_request_commits, project_columns, project_cards, team_members, team_memberships, issue_timeline_events, commit_comment_reactions and issue_comment_reactions since 2.5.0; and commits, contributor_activity, workflow_runs and workflow_jobs since 2.6.0. If a stream fails after an upgrade, check whether it previously skipped that repository or organization.

The same streams also no longer resume part-way through a stream. Previously, a failed attempt resumed from the repository or organization where it stopped. It now restarts from the first one. No data is lost because the retry rereads earlier records. However, a late failure repeats earlier partitions’ requests against your rate limit. The same versions listed above apply. For issues, the connector now saves state once per repository, rather than once per page. An interrupted sync therefore restarts the current repository instead of resuming within it.

contributor_activity fails after five 202 responses

GitHub returns 202 Accepted from the contributor statistics endpoint while calculating the statistics. The connector retries five times at 90-second intervals, as before. Since version 2.6.0, the stream fails if GitHub is still calculating after those retries. Earlier versions logged a message, skipped the repository and reported a successful sync.

Repeated parent records in project_columns, project_cards, commit_comment_reactions and issue_comment_reactions

Since version 2.5.0, these four streams keep a cursor for each parent record. The parent can be a project, column or comment. Each incremental sync re-emits one record per parent whose cursor matches the saved value, for the same reason described above. Existing connections retain their saved cursors. The first sync after upgrading converts the nested state written by earlier versions. Since version 2.6.0, the same behaviour applies to commits, with one cursor per branch, and to workflow_runs and workflow_jobs.

Repeated repository records in comments, issues and review_comments

Since version 2.4.0, each incremental sync of these streams re-emits one record. Its updated_at exactly matches the timestamp where the previous sync stopped. GitHub’s since filter is inclusive, so this record always appears on the first page. The connector no longer removes it locally. The record is unchanged. Destinations using Incremental | Append + Deduped collapse it on the primary key. With Incremental | Append, you’ll see one extra row per repository on each sync.

Repeated records in semi-incremental streams

Since version 2.4.0, each incremental sync re-emits one record per repository for these streams: events, pull_requests, commit_comments, issue_milestones, stargazers, projects, issue_events, deployments and workflows. Its cursor value, updated_at, created_at or starred_at, exactly matches the previous sync’s stopping timestamp. These streams read GitHub’s listing and filter it locally. Previously, the connector retained only records newer than the saved cursor. It now retains the boundary record too. The record is unchanged. Destinations using Incremental | Append + Deduped collapse it on the primary key. With Incremental | Append, you’ll see one extra row per repository on each sync.

The same release includes two smaller changes:

  • pull_requests now sets base.repo_id to the base repository’s id. Earlier versions always emitted null because of a record transformation bug. head.repo_id is unchanged.
  • pull_requests now always lists repositories newest first. Earlier versions listed them oldest first during a connection’s first sync. This allowed a failed first sync to resume part-way through a repository. A failed attempt now restarts the repository, matching every later sync.

review_comments page size

Since version 2.8.2, review_comments requests 100 comments per page instead of 10. On very large repositories, GitHub could time out with the smaller page size and return HTTP 500. The larger page size reduces requests, but each request returns more data and takes longer. This stream’s sync speed may therefore differ from earlier versions. Other large streams still request 10 records per page.

Permissions and scopes

With OAuth authentication, the OAuth2.0 application requests these scopes: repo, read:org, read:repo_hook, read:user, read:discussion, read:project and workflow. For personal access token, manually select the required scopes.

Your token needs at least the repo scope. The user who generates it needs additional permissions for some streams:

  • To sync Collaborators, the user generating the personal access token must be a collaborator. An owner must invite them. If there are no collaborators, the connector syncs no records. Read more about access permissions here.
  • Syncing Teams is available only to authenticated members of a team’s organization. Personal user accounts and their repositories can’t access Teams features. In this case, the connector syncs no records.
  • To sync the Projects stream, enable the Projects feature for the repository.

Troubleshooting

  • Find common troubleshooting guidance for the GitHub source connector in the Airbyte Forum

Changelog

Expand to review
VersionDatePull RequestSubject
2.8.62026-11-1091494Update dependencies
2.8.52026-11-0490970Shorten manifest comments and move the design rationale to the connector docs; no behavior change
2.8.42026-11-0390166Retry GitHub 5xx errors and timeouts with exponential backoff instead of waiting for the rate-limit reset
2.8.32026-11-0390557Update dependencies
2.8.22026-10-2890165Request review_comments with per_page=100 to avoid GitHub timeouts on very large repositories
2.8.12026-10-2789848Update dependencies
2.8.02026-10-2185378Add the workflow_run_attempts stream: every attempt of a workflow run, keyed on [id, run_attempt], since workflow_runs only ever returns the latest attempt of each run
2.7.92026-10-2089187Update dependencies
2.7.82026-10-1388163When GitHub refuses to list a repository’s stargazers because the token is not an admin or collaborator, the stargazers stream now says so in the skip message, and names the skipped repository, instead of reporting a missing repository or missing token scopes. Upgrades to CDK 7.37.0
2.7.72026-10-1388598Update dependencies
2.7.62026-10-1287443Stop reading the releases stream at the first already-synced release instead of the full release history on every sync. Records are unchanged
2.7.52026-10-0687862Update dependencies
2.7.42026-09-2986932Replace custom extractors of reviews, issue_reactions and pull_request_comment_reactions and the custom workflow_runs paginator with airbyte-cdk 7.31.0 features. Records are unchanged
2.7.32026-09-2987172Update dependencies
2.7.22026-09-2286497Make the connector manifest-only on the source-declarative-manifest base image; check and the legacy repository/branch config migrations now run from the manifest. Restore the GraphQL page size of reviews, issue_reactions and pull_request_comment_reactions to 100 and stop reading the deprecated page_size_for_large_streams setting, so the six large streams always use 10
2.7.12026-09-2286600Update dependencies
2.7.02026-09-1685822Declarative migration Step 9 - move the releases, projects_v2, pull_request_stats, reviews, issue_reactions and pull_request_comment_reactions streams to the manifest. These streams now reduce the GraphQL page size and retry the same page when GitHub answers 502/504, instead of retrying the identical oversized query and eventually failing the stream; lowering “Page size for large streams” by hand is no longer needed to get past a timeout. Page-size reduction is also no longer shared between repositories, so one repository timing out no longer changes the page size used for the others. pull_request_stats now always reads newest-first and stops at the first already-synced pull request, rather than only doing so on syncs that already had state. Each incremental sync of the six streams now re-emits the one record per repository whose cursor equals the previous sync’s cursor. A 502/504 on the GraphQL endpoint is now answered by reducing the page size and then, once the page size cannot be reduced any further, by retrying the same page: seven requests over about two minutes at the default page size, and four requests for a source configured with a page size of 1, which has no reduction to make. A GraphQL response that reports its error in the body rather than in the status code is retried again, as it was before 2.7.0, and a repository the token cannot read — which GitHub reports as a body error on an HTTP 200, not as a 404 — is now skipped with a warning like everywhere else in this connector, instead of failing the stream
2.6.12026-09-1786435Follow GitHub’s Link next URL with its after cursor to fix HTTP 422 Pagination with the page parameter is not supported for large datasets on large issues and other REST streams
2.6.02026-09-1085807Declarative migration Step 8 - move the commits, contributor_activity, workflow_runs and workflow_jobs streams to the manifest. Each incremental sync of the three incremental ones now re-emits the one record per branch or repository whose cursor equals the previous sync’s cursor, contributor_activity fails the stream when GitHub is still computing statistics after five retries instead of skipping the repository, and a repository that keeps returning 502/504 after retries fails the stream instead of being skipped with the sync still reported as successful
2.5.02026-09-0885380Declarative migration Step 7 - move the pull_request_commits, project_columns, project_cards, team_members, team_memberships, issue_timeline_events, commit_comment_reactions and issue_comment_reactions streams to the manifest. The four incremental ones now re-emit the one record per parent whose cursor equals the previous sync’s cursor, and a repository that keeps returning 502/504 after retries fails the stream instead of being skipped with the sync still reported as successful
2.4.12026-09-1586032Update dependencies
2.4.02026-09-0985786Declarative migration Steps 5 and 6 - move the events, pull_requests, commit_comments, issue_milestones, stargazers, projects, issue_events, deployments, workflows, comments, issues and review_comments streams to the manifest. Each incremental sync of these streams now re-emits the one record per repository whose cursor equals the previous sync’s cursor, pull_requests fills base.repo_id (always null before) and always lists newest-first, issues checkpoints once per repository instead of once per page, and a repository that keeps returning 502/504 after retries fails the stream instead of being skipped with the sync still reported as successful
2.3.02026-09-0885746Declarative migration Steps 3 and 4 - move the assignees, branches, collaborators, issue_labels, tags, organizations, teams and users streams to the manifest. For these eight streams, a repository or organization that keeps returning 502/504 after retries now fails the stream instead of being skipped with the sync still reported as successful, and a failed attempt restarts the stream instead of resuming from the repository or organization it stopped at. Also fixes a 2.2.0 regression where a 403 on one listed repository (SAML-protected organizations, most often) failed the sync instead of skipping that repository and syncing the rest
2.2.32026-09-0885494Update dependencies
2.2.22026-09-0185253Fix syncs hanging until the platform heartbeat timeout when a wildcard repository entry (owner/*) names a user account instead of an organization: the org-scoped streams no longer receive an unverified login, and a skipped stream slice can no longer be retried indefinitely
2.2.12026-09-0185244Update dependencies
2.2.02026-08-2781428Declarative migration Step 2 - multi-token auth shared by all streams (now rotates off a rate-limited token instead of waiting for its reset), spec in manifest, declarative Repositories stream, new optional num_workers setting for concurrent partition reads, a request budget matching GitHub’s 900-points/minute secondary rate limit, Max Waiting Time now bounding every rate-limit wait so Test connection fails fast instead of sleeping until the reset, and support for GitHub Enterprise Server instances with rate limiting disabled
2.1.422026-08-2585011Update dependencies
2.1.412026-08-1884569Update dependencies
2.1.402026-08-1183943Update dependencies
2.1.392026-08-0483469Update dependencies
2.1.382026-07-2882893Update dependencies
2.1.372026-07-2181799Update dependencies
2.1.362026-07-1482093Promoted release candidate to GA
2.1.36-rc.12026-07-0880133Switch base class from AbstractSource to YamlDeclarativeSource (Step 1 of declarative migration)
2.1.352026-07-0781440Update dependencies
2.1.342026-06-3081084Update dependencies
2.1.332026-06-2380440Update dependencies
2.1.322026-06-1679802Update dependencies
2.1.312026-06-0879196Upgrade cryptography from 44.0.3 to 46.0.7 to resolve CVE-2026-26007
2.1.302026-06-0979312Update dependencies
2.1.292026-06-0278702Update dependencies
2.1.282026-05-0777847Reduce GraphQL releases query cost (mark as large stream), bound page-size halving at 1, and improve 504 Gateway Timeout error messages
2.1.272026-05-0277685Make parse_response and error-handler helpers defensive against unexpected response shapes (HTML error pages, malformed JSON, missing keys)
2.1.262026-05-0277681Treat 410 “Issues/Projects disabled” responses as skip-the-slice instead of failing the stream
2.1.252026-05-0177690Replaced placeholder and raw-passthrough error messages with actionable text including HTTP status codes, affected resources, and likely causes
2.1.242026-04-2877238Update dependencies
2.1.232026-04-2276922Handle None response in read_records error handler so transport-layer failures (connection errors, timeouts) surface as transient errors instead of AttributeError
2.1.222026-04-2274758Fix rate limit sleep blocking heartbeat; classify rate limit errors as transient; increase default max_waiting_time to 120 minutes to cover GitHub’s hourly rate limit window
2.1.212026-04-2174586Update dependencies
2.1.202026-04-1576275Bump airbyte-cdk to >=7.12.0 to fix OAuth scope=null in consent URL
2.1.192026-04-0876090Fix 403 permission errors misclassified as retryable, causing infinite retry loops
2.1.182026-04-0776124Fix silent error swallowing in exception handlers for ContributorActivity, GithubStreamABC, and Releases streams
2.1.172026-04-0776080Fix remaining NameError references to removed MessageRepresentationAirbyteTracedErrors in ContributorActivity stream
2.1.162026-04-0376038Replace deprecated MessageRepresentationAirbyteTracedErrors with AirbyteTracedException
2.1.152026-03-2775508Add declarative OAuth with oauth_connector_input_specification and granular scopes
2.1.142026-03-0974284Fix heartbeat timeout for pull_request_stats by using descending sort on incremental syncs
2.1.132026-03-0373698feat(source-github): use GraphQL API for Releases stream to bypass 10k REST limit
2.1.122026-03-0374204Update dependencies
2.1.112026-02-2473785Update dependencies
2.1.102026-02-1073061Update dependencies
2.1.92026-01-2772375Update dependencies
2.1.82026-01-2071986Update dependencies
2.1.72026-01-1471400Update dependencies
2.1.62025-12-1870729Update dependencies
2.1.52025-12-0270286Update dependencies
2.1.42025-11-2569887Update dependencies
2.1.32025-11-1869421Update dependencies
2.1.22025-11-1169271Update dependencies
2.1.12025-11-0469002Update dependencies
2.1.02025-10-2968637Update dependencies
2.0.02025-10-2268095Breaking Change: Renames +1 and -1 fields to plus_one and minus_one respectively.
1.9.22025-10-2168332Update dependencies
1.9.12025-10-2068197Promoting release candidate 1.9.1-rc.1 to a main version.
1.9.1-rc.12025-10-1367584Graceful error handling of invalid credentials when running operations
1.9.02025-10-1367708Promoting release candidate 1.9.0-rc.3 to a main version.
1.9.0-rc.32025-10-0967589Fix min time to wait on token rate limits
1.9.0-rc.22025-10-0367026Fix converting datetime in workflows stream
1.9.0-rc.12025-10-0266736Update to airbyte-cdk v^7
1.8.422025-09-3066166Update dependencies
1.8.412025-09-0966065Update dependencies
1.8.402025-08-2365375Update dependencies
1.8.392025-08-1664982Update dependencies
1.8.382025-08-0964580Update dependencies
1.8.372025-08-0264231Update dependencies
1.8.362025-07-2663810Update dependencies
1.8.352025-07-1963520Update dependencies
1.8.342025-07-1263158Update dependencies
1.8.332025-07-0562666Update dependencies
1.8.322025-06-2862166Update dependencies
1.8.312025-06-2562054Fix problem with contributor_activity stream when author is not present/none
1.8.302025-06-2361742Handle conflict when empty repositories, we will ignore
1.8.292025-06-2161857Update dependencies
1.8.282025-06-1561603Update dependencies
1.8.272025-06-0754931Update dependencies
1.8.262025-02-2254404Update dependencies
1.8.252025-02-1553703Update dependencies
1.8.242025-02-0152875Update dependencies
1.8.232025-01-2552364Update dependencies
1.8.222025-01-1851666Update dependencies
1.8.212025-01-1151130Update dependencies
1.8.202025-01-0450517Update dependencies
1.8.192024-12-2150055Update dependencies
1.8.182024-12-1449178Update dependencies
1.8.172024-11-2548631Starting with this version, the Docker image is now rootless. Please note that this and future versions will not be compatible with Airbyte versions earlier than 0.64
1.8.162024-11-0548318Update dependencies
1.8.152024-10-2847051Update dependencies
1.8.142024-10-1246766Update dependencies
1.8.132024-10-0546415Update dependencies
1.8.122024-09-2846117Update dependencies
1.8.112024-09-2145742Update dependencies
1.8.102024-09-1445557Update dependencies
1.8.92024-09-0745320Update dependencies
1.8.82024-08-2344592Fix state handling for stream WorkflowRuns
1.8.72024-08-3145061Update dependencies
1.8.62024-08-2444703Update dependencies
1.8.52024-08-1744227Update dependencies
1.8.42024-08-1243749Update dependencies
1.8.32024-08-1042671Update dependencies
1.8.22024-08-2042966Bump cdk version and enable RFR for all non-incremental streams
1.8.12024-07-2042342Update dependencies
1.8.02024-07-1641677Update to 3.4.0 CDK
1.7.132024-07-1341746Update dependencies
1.7.122024-07-1041354Update dependencies
1.7.112024-07-0941221Update dependencies
1.7.102024-07-0641000Update dependencies
1.7.92024-06-2540289Update dependencies
1.7.82024-06-2240128Update dependencies
1.7.72024-06-1739513Update deprecated state handling method
1.7.62024-06-0439078[autopull] Upgrade base image to v1.2.1
1.7.52024-05-2938341Add max_waiting_time to configuration
1.7.42024-05-2138341Update CDK authenticator package
1.7.32024-05-2038299Fixed spec typo
1.7.22024-04-1936636Updating to 0.80.0 CDK
1.7.12024-04-1236636schema descriptions
1.7.02024-03-1936267Pin airbyte-cdk version to ^0
1.6.52024-03-1235986Handle rate limit exception as config error
1.6.42024-03-0835915Fix per stream error handler; Make use the latest CDK version
1.6.32024-02-1535271Update branches schema
1.6.22024-02-1234933Update Airbyte CDK for integration tests
1.6.12024-02-0935087Manage dependencies with Poetry.
1.6.02024-02-0234700Continue Sync on Stream failure
1.5.72024-01-2934598Fix MultipleToken sleep time
1.5.62024-01-2634503Fix MultipleToken rotation logic
1.5.52023-12-2633783Fix retry for 504 error in GraphQL based streams
1.5.42023-11-2032679Return AirbyteMessage if max retry exeeded for 202 status code
1.5.32023-10-2331702Base image migration: remove Dockerfile and use the python-connector-base image
1.5.22023-10-1331386Handle ContributorActivity continuous ACCEPTED response
1.5.12023-10-1231307Increase backoff_time for stream ContributorActivity
1.5.02023-10-1131300Update Schemas: Add date-time format to fields
1.4.62023-10-0431056Migrate spec properties’ repository and branch type to <array>
1.4.52023-10-0231023Increase backoff for stream Contributor Activity
1.4.42023-10-0230971Mark start_date as optional.
1.4.32023-10-0230979Fetch archived records in Project Cards
1.4.22023-09-3030927Provide actionable user error messages
1.4.12023-09-3030839Update CDK to Latest version
1.4.02023-09-2930823Add new stream issue Timeline Events
1.3.12023-09-2830824Handle empty response in stream ContributorActivity
1.3.02023-09-2530731Add new stream ProjectsV2
1.2.12023-09-2230693Handle 404 error in TeamMemberShips
1.2.02023-09-2230647Add support for self-hosted GitHub instances
1.1.12023-09-2130654Rewrite source connection error messages
1.1.02023-08-0330615Add new stream Contributor Activity
1.0.42023-08-0329031Reverted advancedAuth spec changes
1.0.32023-08-0128910Updated advancedAuth broken references
1.0.22023-07-1128144Add archived_at property to Organizations schema parameter
1.0.12023-05-2225838Deprecate “page size” input parameter
1.0.02023-05-1925778Improve repo(s) name validation on UI
0.5.02023-05-1625793Implement client-side throttling of requests
0.4.112023-05-1226025Added more transparent depiction of the personal access token expired
0.4.102023-05-1526075Add more specific error message description for no repos case.
0.4.92023-05-0124523Add undeclared columns to spec
0.4.82023-04-1925312Fix repo name validation
0.4.72023-03-2424457Add validation and transformation for repositories config
0.4.62023-03-2424398Fix caching for get_starting_point in stream “Commits”
0.4.52023-03-2324417Add pattern_descriptors to fields with an expected format
0.4.42023-03-1724255Add field groups and titles to improve display of connector setup form
0.4.32023-03-0422993Specified date formatting in specification
0.4.22023-03-0323467Added user friendly messages, added AirbyteTracedException config_error, updated SAT
0.4.12023-01-2722039Set AvailabilityStrategy for streams explicitly to None
0.4.02023-01-2021457Use GraphQL for issue_reactions stream
0.3.122023-01-1821481Handle 502 Bad Gateway error with proper log message
0.3.112023-01-0621084Raise Error if no organizations or repos are available during read
0.3.102022-12-1520523Revert changes from 0.3.9
0.3.92022-12-1419978Update CDK dependency; move custom HTTPError handling into AvailabilityStrategy classes
0.3.82022-11-1019299Fix events and workflow_runs datetimes
0.3.72022-10-2018213Skip retry on HTTP 200
0.3.62022-10-1117852Use default behaviour, retry on 429 and all 5XX errors
0.3.52022-10-0717715Improve 502 handling for comments stream
0.3.42022-10-0417555Skip repository if got HTTP 500 for WorkflowRuns stream
0.3.32022-09-2817287Fix problem with “null” cursor_field for WorkflowJobs stream
0.3.22022-09-2817304Migrate to per-stream state.
0.3.12022-09-2116947Improve error logging when handling HTTP 500 error
0.3.02022-09-0916534Add new stream WorkflowJobs
0.2.462022-08-1715730Validate input organizations and repositories
0.2.452022-08-1115420”User” object can be “null”
0.2.442022-08-0114795Use GraphQL for pull_request_comment_reactions stream
0.2.432022-07-2615049Bugfix schemas for streams deployments, workflow_runs, teams
0.2.422022-07-1214613Improve schema for stream pull_request_commits added “null”
0.2.412022-07-0314376Add Retry for GraphQL API Resource limitations
0.2.402022-07-0114338Revert: “Rename field mergeable to is_mergeable”
0.2.392022-06-3014274Rename field mergeable to is_mergeable
0.2.382022-06-2713989Use GraphQL for reviews stream
0.2.372022-06-2113955Fix “secondary rate limit” not retrying
0.2.362022-06-2013926Break point added for workflows_runs stream
0.2.352022-06-1613763Use GraphQL for pull_request_stats stream
0.2.342022-06-1413707Fix API sorting, fix get_starting_point caching
0.2.332022-06-0813558Enable caching only for parent streams
0.2.322022-06-0713531Fix different result from get_starting_point when reading by pages
0.2.312022-05-2413115Add incremental support for streams WorkflowRuns
0.2.302022-05-0912294Add incremental support for streams CommitCommentReactions, IssueCommentReactions, IssueReactions, PullRequestCommentReactions, Repositories, Workflows
0.2.292022-05-0412482Update input configuration copy
0.2.282022-04-2111893Add new streams TeamMembers, TeamMemberships
0.2.272022-04-0211678Fix “PAT Credentials” in spec
0.2.262022-03-3111623Re-factored incremental sync for Reviews stream
0.2.252022-03-3111567Improve code for better error handling
0.2.242022-03-309251Add Streams Workflow and WorkflowRuns
0.2.232022-03-1711212Improve documentation and spec for Beta
0.2.222022-03-1010878Fix error handling for unavailable streams with 404 status code
0.2.212022-03-0410749Add new stream ProjectCards
0.2.202022-02-1610385Add new stream Deployments, ProjectColumns, PullRequestCommits
0.2.192022-02-0710211Add human-readable error in case of incorrect organization or repo name
0.2.182021-02-0910193Add handling secondary rate limits
0.2.172021-02-029999Remove BAD_GATEWAY code from backoff_time
0.2.162021-02-029868Add log message for streams that are restricted for OAuth. Update oauth scopes.
0.2.152021-01-269802Add missing fields for auto_merge in pull request stream
0.2.142021-01-219664Add custom pagination size for large streams
0.2.132021-01-209619Fix logging for function should_retry
0.2.112021-01-179492Remove optional parameter Accept for reaction`s streams to fix error with 502 HTTP status code in response
0.2.102021-01-037250Use CDK caching and convert PR-related streams to incremental
0.2.92021-12-299179Use default retry delays on server error responses
0.2.82021-12-078524Update connector fields title/description
0.2.72021-12-068518Add connection retry with GitHub
0.2.62021-11-248030Support start date property for PullRequestStats and Reviews streams
0.2.52021-11-218170Fix slow check connection for organizations with a lot of repos
0.2.42021-11-117856Resolve $ref fields in some stream schemas
0.2.32021-10-066833Fix config backward compatability
0.2.22021-10-056761Add oauth worflow specification
0.2.12021-09-226223Add option to pull commits from user-specified branches
0.2.02021-09-195898 and 6227Don’t minimize any output fields & add better error handling
0.1.112021-09-155949Add caching for all streams
0.1.102021-09-095860Add reaction streams
0.1.92021-09-025788Handling empty repository, check method using RepositoryStats stream
0.1.82021-09-015757Add more streams
0.1.72021-08-275696Handle negative backoff values
0.1.62021-08-185223Add MultipleTokenAuthenticator
0.1.52021-08-185456Fix set up validation
0.1.42021-08-135136Support syncing multiple repositories/organizations
0.1.32021-08-035156Extended existing schemas with users property for certain streams
0.1.22021-07-134708Fix bug with IssueEvents stream and add handling for rate limiting
0.1.12021-07-074590Fix schema in the pull_request stream
0.1.02021-07-064174New Source: GitHub
Advanced Analytica

Most businesses are sitting on information they can’t use at speed. The people who hold it are busy, and when they leave, the knowledge goes with them. The documents are static, and the processes around them weren’t designed for AI.

Advanced Analytica works with these businesses to turn that information into structured data assets. Once they're integrated, AI systems can govern it, and the business can use it. We bring in the strategy, the specification and the AI skills to rework how the business operates.

MOVE FAST. STAY SAFE.