Frequently asked questions

Odara ETL: questions and straight answers

Including the uncomfortable ones. If something here is out of date, write to [email protected].

What is Odara ETL?

A data-integration (ETL/ELT) platform written in Rust. One native binary holds the engine, the API and a browser-based visual editor. SQL steps run on Apache DataFusion, Python steps run next to them, and data moves between steps as Apache Arrow. Each environment (for example DEV, STG, PROD) decides where its runs happen: on the Odara server, in a Docker container, on AWS ECS Fargate or on Google Cloud Run.

Is Odara ETL free? Are there limits?

Yes, it is free of charge for any use, including commercial use. Since version 0.4.3 there is no limit on the number of users, and there are no quotas on pipelines, rows, schedules or machines. There is no licence key and no activation: the installed server does not contact odara.rs.

Is Odara ETL open source?

No. It is proprietary software, free of charge. Its licence includes a continuity pledge: if Odara ETL stops being maintained, the source code of its last release will be published under the Apache License 2.0, so anyone can keep running, fixing and building it.

What happens to my work if Odara stops being maintained?

Three things protect it. Your pipelines are stored as JSON files that can be versioned in Git, and your SQL and Python steps are plain code you can read without Odara. The installed software keeps working with no licence server to switch off. And the continuity pledge: the source code of the last release is published under the Apache License 2.0 after a public announcement that development has ended, or twelve consecutive months with no new release and no answer to [email protected].

Who makes Odara ETL?

An independent developer based in Europe, not a company. Odara is young (version 0.4 in November 2026) and released often; every release is listed in the release notes with what changed and how it was tested.

Is it ready for production?

Odara is pre-1.0. Each release is tested end to end before it is published, including runs on real AWS and Google Cloud accounts and installation tests on Linux and Windows. Expect changes between 0.x versions: test upgrades on a copy first and read the release notes.

Which systems does it run on?

Linux x86_64 (.deb, .rpm, tar.gz) and Windows 64-bit. No macOS or ARM64 build yet. The editor runs in any modern browser. Python steps need Python 3.9 or later on the machine; Docker is only needed for isolated runs in Docker.

What can it connect to?

71 node types. Databases: PostgreSQL, MySQL, Oracle, SQL Server, MongoDB, SQLite and ODBC. Warehouses: Snowflake, BigQuery, Redshift, ClickHouse and DuckDB. Files: CSV, Excel, XML, JSON, NDJSON and Parquet. Storage: S3, Azure Blob, Google Drive, FTP/SFTP and local files. REST APIs as source and target, email, and control-flow nodes. Streaming sources (Kafka, RabbitMQ, NATS, MQTT) are not in the current build.

How do SQL and Python work together?

SQL steps run inside the engine on Apache DataFusion. A Python step runs in its own process and receives its input as Apache Arrow (with pandas and PyArrow available), so data is never converted to rows or text between steps.

Where does my data go?

Only where your pipelines send it. Odara runs on your machines or in your own AWS or Google Cloud account; nothing goes to odara.rs. In isolated runs, secret values never travel in the run package: on AWS and Google each run gets its own vault entry, deleted when the run ends.

Does it use AI? Does it send my data to an AI provider?

The AI assistant is optional and uses your own key: Anthropic, Google Gemini, DeepSeek or any OpenAI-compatible endpoint. Nothing is sent to a provider unless you configure one and use the assistant.

Can it replace Talend Open Studio?

That is what it is built for. Its editor works like a Talend job (sources, a Mapper like tMap, SQL and Python steps, targets, and Maestro to orchestrate jobs). The Talend Import Utility (Beta) reads a Talend project export, reports every job and component, explains each job with your AI provider, and proposes Odara pipelines for you to review. The Run External Job node keeps exported Talend jobs running while you migrate.

How do I get help?

The documentation, the Academy (free courses and step-by-step guides), the Community forum on odara.rs, and email at [email protected].

More: the licence, the documentation, Odara and Apache Hop, and leaving Talend Open Studio.