Skip to main content

Language SDKs

Conduit connectors and processors speak a stable gRPC protocol (processors can also compile to WASM), so plugins are not limited to one language. Six languages have official SDKs, maintained in two tiers. Every other language can build against the same open protocol plus a conformance kit.

Python — the next SDK (pre-alpha)​

If the system you are integrating already has a mature Python client — a database driver, boto3, httpx, a vector-store SDK — you can reach for it directly instead of reimplementing it. That makes Python a natural fit for the data- and AI-adjacent edges of a pipeline.

The Python connector SDK is pre-alpha and in active development: the API is real and usable today, but nothing has shipped to PyPI yet. The distribution is conduitio; the import package is conduit, exposing two namespaces:

  • conduit.connector — the connector SDK. Subclass Source or Destination and call serve(...) to run a standalone plugin.
  • conduit.client — a gRPC control client for driving a running Conduit instance.

Write a connector in Python →

Full tier — Go, Python, TypeScript, Rust​

Full-tier languages get connectors over gRPC, processors (WASM in-process and/or gRPC out-of-process), the processor state API, and an embedded client.

Enterprise tier — Java, C#​

Enterprise-tier languages get connectors over gRPC and processors on the gRPC out-of-process runtime (including the state API), with an embedded client generated from gRPC bindings.

  • Java (roadmap) — the native path for Kafka Connect connector authors. Existing Kafka Connect JARs already run through the Kafka Connect connector.
  • C# (roadmap) — gRPC connectors, processors, and the state API.

Every other language​

Community-tier languages build against the same open protocol, with a conformance kit. SDKs that pass the kit are listed as community-maintained.

The full plan — phases, parity, and the two processor runtimes — is on the roadmap.