Do not miss the new connectors we support!
You can request new connectors directly here.
Note: Airbyte is not built on top of Singer, but is compatible with Singer's protocol. Airbyte's ambitions go beyond what Singer enables to do, so we are building our own protocol that will keep its compatibility with Singer's one.
Check out our connector roadmap to see what we're currently working on.

New sources:

New sources:

New source:
New destinations:

New source:

New destinations:

New sources:
New features:
Exchange Rates Source: add ignore_weekends option. Facebook Source: add the videos stream. Freshdesk Source: removed the limitation in streams pagination. Jira Source: add option to render fields in HTML format. MongoDB v2 Source: improve read performance. Pipedrive Source: specify schema for "persons" stream. PostgreSQL Source: exclude tables on which user doesn't have select privileges. SurveyMonkey Source: improve connection check.

New destination:
New sources:
New features:
MySQL Source: Now has basic performance tests. Salesforce Source: We now automatically transform and handle incorrect data for the anyType and calculated types.

New destinations:
New sources:
New features:
Shopify Source: Now supports the FulfillmentOrders and Fulfillments streams. Greenhouse Source: Now supports the Demographics stream. Recharge Source: Broken requests should now be re-requested with improved backoff. Stripe Source: Now supports the checkout_sessions, checkout_sessions_line_item, and promotion_codes streams. Db2 Source: Now supports SSL.

New destination:
New sources:
New features:
BigQuery Destination: You can now run transformations in batches, preventing queries from hitting BigQuery limits. (contributed by @Andrés Bravo) S3 Source: Memory and Performance optimizations, also some fancy new PyArrow CSV configuration options. Zuora Source: Now supports Unlimited as an option for the Data Query Live API. Clickhouse Source: Now supports SSL and connection via SSH tunneling.

New source:
New feature: MSSQL destination: Now supports basic normalization

New sources:
New features:
MySQL destination: Now supports connection via TLS/SSL BigQuery (denormalized) destination: Supports reading BigQuery types such as date by reading the format field (contributed by @Nicolas Moreau) Hubspot source: Added contacts associations to the deals stream. GitHub source: Now supports pulling commits from user-specified branches. Google Search Console source: Now accepts admin email as input when using a service account key. Greenhouse source: Now identifies API streams it has access to if permissions are limited. Marketo source: Now Airbyte native. S3 source: Now supports any source that conforms to the S3 protocol (Non-AWS S3). Shopify source: Now reports pre_tax_price on the line_items stream if you have Shopify Plus. Stripe source: Now actually uses the mandatory start_date config field for incremental syncs.

New destinations:
New sources:
New features:
Google Ads source: You can now specify user-specified queries in GAQL. GitHub source: All streams with a parent stream use cached parent stream data when possible. Shopify source: Substantial performance improvements to the incremental sync mode. Stripe source: Now supports the PaymentIntents stream. Pipedrive source: Now supports the Organizations stream. Sendgrid source: Now supports the SingleSendStats stream. Bing Ads source: Now supports the Report stream. GitHub source: Now supports the Reactions stream. MongoDB source: Now Airbyte native!

New source:
New destinations:
New features:
  • S3 source: Support for Parquet format.
  • Github source: Branches, repositories, organization users, tags, and pull request stats streams added (contributed by @Christopher Wu).
  • BigQuery destination: Added GCS upload option.
  • Salesforce source: Now Airbyte native.
  • Redshift destination: Optimized for performance.
Bug fixes:
  • Pipedrive source: Output schemas no longer remove timestamp from fields.
  • Github source: Empty repos and negative backoff values are now handled correctly.
  • Harvest source: Normalization now works as expected.
  • All CDC sources: Removed sleep logic which caused exceptions when loading data from high-volume sources.
  • Slack source: Increased number of retries to tolerate flaky retry wait times on the API side.
  • Slack source: Sync operations no longer hang indefinitely.
  • Jira source: Now uses updated time as the cursor field for incremental sync instead of the created time.
  • Intercom source: Fixed inconsistency between schema and output data.
  • HubSpot source: Streams with the items property now have their schemas fixed.
  • HubSpot source: Empty strings are no longer handled as dates, fixing the deals, companies, and contacts streams.
  • Typeform source: Allows for multiple choices in responses now.
  • Shopify source: The type for the amount field is now fixed in the schema.
  • Postgres destination: \u0000(NULL) value processing is now fixed.

New sources:
Bug fixes:
  • Shopify source: Rate limit throttling fixed.

New source:
New features:
  • GitHub source: Add support for rotating through multiple API tokens.
  • Google Ads source: Added UserLocationReport stream.
  • source: Added the order_items stream.
Bug fixes:
  • Postgres source: Fix out-of-memory issue with CDC interacting with large JSON blobs.
  • Intercom source: Pagination now works as expected.

New source:
New destination:
New features:
  • Chargebee source: Adds support for the items, item prices and attached items endpoints.
Bug fixes:
  • QuickBooks source: Now uses the number data type for decimal fields.
  • HubSpot source: Fixed empty string inside of the number and float datatypes.
  • GitHub source: Validation fixed on non-required fields.
  • BigQuery destination: Now supports processing of arrays of records properly.
  • Oracle destination: Fixed destination check for users without DBA role.

New sources:
New features:
  • Shopify source: The status property is now in the Products stream.
  • Amazon Seller Partner source: Added support for GET_MERCHANT_LISTINGS_ALL_DATA and GET_FBA_INVENTORY_AGED_DATA stream endpoints.
  • GitHub source: Existing streams now don't minify the user property.
  • HubSpot source: Updated user-defined custom field schema generation.
  • Zendesk source: Migrated from Singer to the Airbyte CDK.
  • Amazon Seller Partner source: Migrated to the Airbyte CDK.
Bug fixes:
  • HubSpot source: Casting exceptions are now logged correctly.
  • S3 source: Fixed bug where syncs could hang indefinitely.
  • Shopify source: Fixed the products schema to be in accordance with the API.
  • PayPal Transactions source: Fixed the start date minimum to be 3 years rather than 45 days.
  • Google Ads source: Added the login-customer-id setting.
  • Intercom source: Rate limit corrected from 1000 requests/minute from 1000 requests/hour.
  • S3 source: Fixed bug in spec to properly display the format field in the UI.
New CDK features:
  • Now allows for setting request data in non-JSON formats.

New sources:
Bug fixes:
  • Shopify source: Fixed the products schema to be in accordance with the API.
  • BigQuery source: No longer fails with Array of Records data types.
  • BigQuery destination: Improved logging, Job IDs are now filled with location and Project IDs.

New sources:
New destinations:
New Features:
  • Java Connectors: Now have config validators for check, discover, read, and write calls
  • Stripe source: All subscription types are returnable (including expired and canceled ones).
  • Mixpanel source: Migrated to the CDK.
  • Intercom source: Migrated to the CDK.
  • Google Ads source: Now supports the Campaigns, Ads, AdGroups, and Accounts streams.
Bug Fixes:
  • Facebook source: Improved rate limit management
  • Instagram source: Now supports old format for state and automatically updates it to the new format.
  • Sendgrid source: Now gracefully handles malformed responses from API.
  • Jira source: Fixed dbt failing to normalize schema for the labels stream.
  • MySQL destination: Does not fail anymore with columns that contain JSON data.
  • Slack source: Now does not fail stream slicing on reading threads.

3 new sources:
New Features:
  • MySQL destination: Now supports normalization!
  • MSSQL source: Now supports CDC (Change Data Capture).
  • Snowflake destination: Data coming from Airbyte is now identifiable.
  • GitHub source: Now handles rate limiting.
Bug Fixes:
  • GitHub source: Now uses the correct cursor field for the IssueEvents stream.
  • Square source: send_request method is no longer broken.

7 new sources:
New Features:
  • S3 destination: Now supports anyOf, oneOf and allOf schema fields.
  • Instagram source: Migrated to the CDK and has improved error handling.
  • Snowflake source: Now has comprehensive data type tests.
  • Shopify source: Change the default stream cursor field to update_at where possible.
  • Shopify source: Add support for draft orders.
  • MySQL destination: Now supports normalization.
Connector Development:
  • Python CDK: Now allows setting of network adapter args on outgoing HTTP requests.
  • Abstract classes for non-JDBC relational database sources.
  • GitHub source: Fixed issue with locked breaking normalization of the pull_request stream.
  • PostgreSQL source: Fixed decimal handling with CDC.
  • Okta source: Fix endless loop when syncing data from logs stream.

  • Looker source: Now supports the Run Look stream.
  • Google Adwords: CI is fixed and new version is published.
  • Slack source: Now Airbyte native and supports channels, channel members, messages, users, and threads streams.
  • Freshdesk source: Does not fail after 300 pages anymore.
  • MSSQL source: Now has comprehensive data type tests.

1 new source:
New features:
  • S3 destination: supports Avro and Jsonl output!
  • BigQuery destination: now supports loading JSON data as structured data.
  • Looker source: Now supports self-hosted instances.
  • Facebook source: is now migrated to the CDK.

1 new source:
New features:
  • Postgres source: now has comprehensive data type tests.
  • Google Ads source: now uses the Google Ads Query Language!
  • S3 destination: supports Parquet output!
  • S3 destination: supports Minio S3!
  • BigQuery destination: credentials are now optional.

1 new destination:
3 new sources:
New features:
  • Jira source: now supports all available entities in Jira Cloud.
  • ExchangeRatesAPI source: clearer messages around unsupported currencies.
  • MySQL source: Comprehensive core extension to be more compatible with other JDBC sources.
  • BigQuery destination: Add dataset location.
  • Shopify source: Add order risks + new attributes to orders schema for native connector
  • MSSQL destination: fixed handling of unicode symbols.
Connector development updates:
  • Containerized connector code generator.
  • Added JDBC source connector bootstrap template.
  • Added Java destination generator.

2 new sources:
New features:
  • MySQL CDC now only polls for 5 minutes if we haven't received any records (#3789)
  • Python CDK now supports Python 3.7.X (#3692)
  • File source: now supports Azure Blob Storage (#3660)
  • Recurly source: now uses type number instead of integer (#3769)
  • Stripe source: fix types in schema (#3744)
  • Stripe source: output number instead of int (#3728)
  • MSSQL destination: fix issue with unicode symbols handling (#3671)

4 new sources:
Progress on connectors:
  • CDC for MySQL is now available!
  • Sendgrid source: support incremental sync, as rewritten using HTTP CDK (#3445)
  • Github source bugfix: exception when parsing null date values, use created_at as cursor value for issue_milestones (#3314)
  • Slack source bugfix: don't overwrite thread_ts in threads stream (#3483)
  • Facebook Marketing source: allow configuring insights lookback window (#3396)
  • Freshdesk source: fix discovery (#3591)

1 new destination: MSSQL
1 new source: ClickHouse
Progress on connectors:
  • Shopify: make this source more resilient to timeouts (#3409)
  • Freshdesk bugfix: output correct schema for various streams (#3376)
  • Iterable: update to use latest version of CDK (#3378)

1 new destination: MySQL
2 new sources:
Progress on connectors:
  • Zoom: bugfix on declaring correct types to match data coming from API (#3159), thanks to vovavovavovavova
  • Smartsheets: bugfix on gracefully handling empty cell values (#3337), thanks to Nathan Nowack
  • Stripe: fix date property name, only add connected account header when set, and set primary key (#3210), thanks to Nathan Yergler

2 new sources:
Progress on connectors:
  • Appstore: bugfix private key handling in the UI (#3201)
  • Facebook marketing: Wait longer (5 min) for async jobs to start (#3116), thanks to Max Krog
  • Stripe: support reading data from connected accounts (#3121), and 2 new streams with Refunds & Bank Accounts (#3030) (#3086)
  • Redshift destination: Ignore records that are too big (instead of failing) (#2988)
  • MongoDB: add supporting TLS and Replica Sets (#3111)
  • HTTP sources: bugfix on handling array responses gracefully (#3008)

  • Zendesk Talk: fix normalization failure (#3022), thanks to yevhenii-ldv
  • Github: pull_requests stream only incremental syncs (#2886) (#3009), thanks to Zirochkaa
  • Create streaming writes to a file and manage the issuance of copy commands for the destination (#2921)
  • Redshift: make Redshift part size configurable. (#3053)
  • HubSpot: fix argument error in log call (#3087) (#3087) , thanks to Nathan Yergler

3 new source connectors!
Other progress on connectors:
  • Postgres source/destination: add SSL option, thanks to Marcos Marx (#2757)
  • Google sheets bugfix: handle duplicate sheet headers, thanks to Aneesh Makala (#2905)
  • Source Google Adwords: support specifying the lookback window for conversions, thanks to Harshith Mullapudi (#2918)
  • MongoDB improvement: speed up mongodb schema discovery, thanks to Yury Koleda (#2851)
  • MySQL bugfix: parsing Mysql jdbc params, thanks to Vasily Safronov (#2891)
  • CSV bugfix: discovery takes too much memory (#2089)
  • A lot of work was done on improving the standard tests for the connectors, for better standardization and maintenance!

  • New connector: Google Workspace Admin Reports (audit logs)
  • Bugfix in the base python connector library that caused errors to be silently skipped rather than failing the sync
  • bugfix: to point to the updated API URL
  • Redshift destination bugfix: quote keywords “DATETIME” and “TIME” when used as identifiers
  • GitHub bugfix: syncs failing when a personal repository doesn’t contain collaborators or team streams available
  • Mixpanel connector: sync at most the last 90 days of data in the annotations stream to adhere to API limits

  • We started measuring throughput of connectors. This will help us improve that point for all connectors.
  • Redshift: implemented Copy strategy to improve its throughput.
  • Instagram: bugfix an issue which caused media and media_insights streams to stop syncing prematurely.
  • Support NCHAR and NVCHAR types in SQL-based database sources.
  • Add the ability to specify a custom JDBC parameters for the MySQL source connector.

  • 2 new source connectors: Gitlab and Airbyte-native HubSpot
  • Developing connectors now requires almost no interaction with Gradle, Airbyte’s monorepo build tool. If you’re building a Python connector, you never have to worry about developing outside your typical flow. See the updated documentation.

  • 2 new source connectors: Instagram and Google Directory
  • Facebook Marketing: support of API v10
  • Google Analytics: support incremental sync
  • Jira: bug fix to consistently pull all tickets
  • HTTP Source: bug fix to correctly parse JSON responses consistently

  • 1 new source connector: MongoDB
  • Google Analytics: Support chunked syncs to avoid sampling
  • AppStore: fix bug where the catalog was displayed incorrectly

  • New native HubSpot connector with schema folder populated
  • Facebook Marketing connector: add option to include deleted records

  • Bug fixes:
    • Google Analytics: add the ability to sync custom reports
    • Apple Appstore: bug fix to correctly run incremental syncs
    • Exchange rates: UI now correctly validates input date pattern
    • File Source: Support JSONL (newline-delimited JSON) format
    • Freshdesk: Enable controlling how many requests per minute the connector makes to avoid overclocking rate limits

  • 1 new destination connector: MeiliSearch
  • 2 new sources that support incremental append: Freshdesk and Sendgrid
  • Other fixes:
    • Thanks to @ns-admetrics for contributing an upgrade to the Shopify source connector which now provides the landing_site field containing UTM parameters in the Orders table.
    • Sendgrid source connector supports most available endpoints available in the API
    • Facebook Source connector now supports syncing Ad Insights data
    • Freshdesk source connector now supports syncing satisfaction ratings and conversations
    • Microsoft Teams source connector now gracefully handles rate limiting
    • Bug fix in Slack source where the last few records in a sync were sporadically dropped
    • Bug fix in Google Analytics source where the last few records in sync were sporadically dropped
    • In Redshift source, support non alpha-numeric table names
    • Bug fix in Github Source to fix instances where syncs didn’t always fail if there was an error while reading data from the API

  • Sources that we improved reliability for (and that became “certified”):
    • Certified sources: Files and Shopify
    • Enhanced continuous testing for Tempo and Looker sources
  • Other fixes / features:
    • Correctly handle boolean types in the File Source
    • Add docs for App Store source
    • Fix a bug in Snowflake destination where the connector didn’t check for all needed write permissions, causing some syncs to fail

  • Improved reliability with our best practices on : Google Sheets, Google Ads, Marketo, Tempo
  • Support incremental for Facebook and Google Ads
  • The Facebook connector now supports the FB marketing API v9

  • 1 new source: App Store (thanks to @Muriloo)
  • Fixes on connectors:
    • Bug fix writing boolean columns to Redshift
    • Bug fix where getting a connector’s input configuration hung indefinitely
    • Stripe connector now gracefully handles rate limiting from the Stripe API

  • 1 new source: Tempo (thanks to @thomasvl)
  • Incremental support for 3 new source connectors: Salesforce, Slack and Braintree
  • Fixes on connectors:
    • Fix a bug in MSSQL and Redshift source connectors where custom SQL types weren't being handled correctly.
    • Improvement of the Snowflake connector from @hudsondba (batch size and timeout sync)

  • Incremental support for 2 new source connectors: Mixpanel and HubSpot
  • Fixes on connectors:
    • Fixed a bug in the github connector where the connector didn’t verify the provided API token was granted the correct permissions
    • Fixed a bug in the Google sheets connector where rate limits were not always respected
    • Alpha version of Facebook marketing API v9. This connector is a native Airbyte connector (current is Singer based).

New sources: Plaid (contributed by tgiardina), Looker

New sources: Drift, Microsoft Teams