Airflow Summit 2026 is coming August 31 - September 2 in Austin, TX. Register now to secure your spot!

apache-airflow-providers-cohere

The cohere provider gives Dags direct access to Cohere’s own Embed API — this page compares that choice against common.ai.

When to use this provider

Use cohere when a Dag needs Cohere’s native embedding models specifically:

  • CohereEmbeddingOperator — call Cohere’s Embed API directly via CohereHook.

Use apache-airflow-providers-common-ai instead when the embedding step should stay vendor-neutral:

  • Document-to-vector-store pipelines with its document loader, embedding, and retrieval operators (see Common AI Operators), which are not tied to Cohere’s embedding models.

apache-airflow-providers-cohere package

Cohere

Release: 1.6.6

Provider package

This package is for the cohere provider. All classes for this package are included in the airflow.providers.cohere python package.

Installation

You can install this package on top of an existing Airflow installation via pip install apache-airflow-providers-cohere. For the minimum Airflow version supported, see Requirements below.

Requirements

The minimum Apache Airflow version supported by this provider distribution is 2.11.0.

PIP package

Version required

apache-airflow

>=2.11.0

apache-airflow-providers-common-compat

>=1.8.0

cohere

>=5.13.4

fastavro

>=1.10.0; python_version >= "3.13" and python_version < "3.14"

fastavro

>=1.12.1; python_version >= "3.14"

Downloading official packages

You can download officially released packages and verify their checksums and signatures from the Official Apache Download site

Was this entry helpful?