---
id: fivetran-connector-sdk
version: "2.11.0"
license: unclear
license_treatment: permissive
maintenance: active
---

# fivetran-connector-sdk — Build custom connectors on Fivetran platform

License: permissive · Maintenance: active · Downloads: 216.5K/mo

## What it is and what it does

The fivetran-connector-sdk is a framework for writing custom data connectors that execute within Fivetran's managed cloud infrastructure. Instead of building a standalone service or relying on unsupported third-party connectors, you write Python code using the SDK's Connector, Logging, and Operations classes, and Fivetran handles scheduling, compute, and orchestration. The SDK is designed for scenarios where Fivetran lacks a native connector—private APIs, custom applications, unsupported file formats, or data that requires filtering or anonymization before reaching your destination.

The package provides a structured way to define how data is extracted from your source, transformed if needed, and delivered to Fivetran's platform. It depends on grpcio and grpcio-tools for communication, requests for HTTP operations, and utility libraries like tqdm and prompt-toolkit. The framework abstracts away infrastructure concerns, letting you focus on the connector logic itself.

Use it for:

- Extract data from a proprietary internal API that Fivetran doesn't support natively.
- Pre-process or filter sensitive data before it enters your data warehouse.
- Build a connector for a custom file format or database that requires special handling.
- Integrate a private data source without exposing it to a third-party service.
- Anonymize or transform data on ingestion to meet compliance requirements.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Build and deploy custom Python data connectors that run in Fivetran's cloud environment, handling data extraction, transformation, and loading for sources Fivetran doesn't natively support.

Yes, if you need to ingest data from a source Fivetran doesn't support and want to avoid building and hosting a separate connector service. The SDK is actively maintained, has low install friction, carries a permissive license, and integrates directly with Fivetran's platform. Install only if you have a specific unsupported source in mind and are comfortable writing Python code to define the connector logic.

## Install

pip install fivetran-connector-sdk
uv add fivetran-connector-sdk
poetry add fivetran-connector-sdk

## Installing fivetran-connector-sdk

Before you install: Low friction installation with a pure-Python wheel. Active maintenance as of 2026-08-14 with recent releases and an actively maintained repository. Requires Python ≥3.10 and depends on grpcio, grpcio-tools, and other standard libraries.

License in practice: MIT license (permissive) means you can use, modify, and distribute connectors built with this SDK with minimal legal constraints.

Quickstart:

pip install fivetran-connector-sdk

from fivetran_connector_sdk import Connector
from fivetran_connector_sdk import Logging as log
from fivetran_connector_sdk import Operations as op

Requires Python ≥3.10 and ≤3.14; Windows 10+ (64-bit), macOS 13+, or Linux distributions such as Ubuntu 20.04+ (arm64 or x86_64).

Verify before relying:

- Whether the SDK supports incremental sync, CDC, or only full refresh patterns
- Performance characteristics and data volume limits for connectors built with this SDK
- Availability of pre-built connector examples beyond what's mentioned in the GitHub repository

## Package facts

- License: not declared (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 216.5K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

custom data connector framework, fivetran connector development, etl connector sdk, build custom data pipeline, fivetran integration builder, python data connector, custom source connector, data-integration, etl-framework

[View on SkillFed](https://skillfed.io/packages/fivetran-connector-sdk) · [View on PyPI](https://pypi.org/project/fivetran-connector-sdk/)