Airbyte

Airbyte is an open-source data integration platform for data engineering teams that syncs apps, databases and files into warehouses, lakes and AI apps.

Airbyte keeps a warehouse, a lake or an AI application fed from every source a company runs, without a connector written by hand for each one. Its catalog of connectors is configured rather than built, which removes the work of writing and maintaining one per source. Airbyte, Inc. of San Francisco makes and owns it; the company started the project as open source in 2020 and remains independent. The platform is organised around sources, destinations and connections. A source pulls from a system such as Salesforce, HubSpot, PayPal, a Kafka topic, an S3 bucket or an SFTP folder. A destination lands the data in a warehouse, a lakehouse or another Kafka topic. A connection joins the two with a sync schedule and tabs for Status, Timeline, Schema, Mappings, Transformation and Settings. Mappings hash, encrypt, rename or filter fields on each stream before the data lands. File sources such as OneDrive and SharePoint parse Markdown, TXT, PDF, Word, PowerPoint and Google documents into text rows for AI workloads. The Connector Setup Assistant, in beta, configures a new source through a chat instead of a form. On data activation destinations, rejected records go to an S3 bucket, are repaired upstream and are processed again on the next sync. On the cloud tiers, a dbt Cloud integration runs dbt transformations in the destination immediately after a sync. PyAirbyte brings the same connectors into Python code. Airbyte suits data engineering and analytics teams that replicate many sources into one warehouse or lake and want to choose where the platform runs. Teams that prefer open source run Airbyte Core on their own servers. Teams that want no infrastructure use the managed cloud. Single sign-on, field renaming and faster sync intervals arrive on the Plus and Pro tiers, and role-based access control and multiple workspaces on Pro. Regulated organisations use Enterprise Flex for on-premises, in-boundary or multi-region deployment, with data planes in their own infrastructure and only metadata reaching Airbyte's managed control plane. It is bought as one platform on one login, billed monthly by credits on Standard and Plus and by capacity on Pro. Airbyte's place in a stack is between the operational systems and the analytics or inference layer. A typical setup lands Salesforce and HubSpot records in a Databricks lakehouse or a warehouse and reads from and writes to Kafka topics. The same workspace pulls PayPal and Calendly data for reporting and loads documents from SharePoint or OneDrive for retrieval by AI applications. Airbyte works beside the warehouse, lakehouse and Python tooling a team already runs, and Airbyte Core keeps the same connectors available inside a team's own network.