Getting started.
What Odara ETL is, its architecture, how to install it on Linux or Windows, your first pipeline and the system requirements.
1. Getting Started
1.1 What is Odara ETL
Odara ETL is a data-integration (ETL/ELT) platform, free to download and use. It lets you design data pipelines visually, read and write databases, warehouses, files, cloud storage and APIs with 70 node types, transform data with SQL or Python, orchestrate pipelines with Maestro and schedule them. Pipelines run on your own server, or — per environment — in an isolated container or in your own AWS or Google Cloud account.
Built in Rust for the backend (no JVM, no GC pauses) and React for the frontend, Odara combines modern engineering with a familiar drag-and-drop workflow inspired by tools like Talend and Informatica.
Visual pipeline editor
Drag, drop, connect. Configure each node with a side panel.
SQL + Python transforms
Use DataFusion SQL or embedded Python (pandas / PyArrow).
70 node types
PostgreSQL, MySQL, Oracle, SQL Server, MongoDB, SQLite, ODBC, Snowflake, BigQuery, Redshift, ClickHouse, DuckDB, CSV, Excel, XML, S3, Azure Blob, Google Drive, FTP/SFTP, REST, Email, and more.
AI assistant
Generate SQL, Python, or full pipelines from natural language, with your own key (DeepSeek, Anthropic, Gemini, OpenAI-compatible).
Maestro orchestration
Combine pipelines into series or parallel workflows with retry policies.
Git integration
Version-control your pipelines directly from the editor.
Runs where you choose
Each environment runs Local, in Docker, on AWS ECS Fargate or on Google Cloud Run (section 15).
Talend Import (Beta)
Reads a Talend project export and proposes Odara pipelines you review (section 16).
1.2 Architecture at a glance
Odara is a single-binary backend that serves the UI and exposes a REST + SSE API. Data flows through pipelines as Apache Arrow RecordBatches, SQL runs in DataFusion, and Python runs in isolated subprocesses with PyArrow for data exchange.
- Backend:
odarax-api(Rust / Axum / Tokio) — serves HTTP API + static UI - Engine:
odarax-core(DataFusion + custom connectors) - Frontend: React + React Flow (visual editor) + Monaco (code editor)
- Storage: SQLite per project (WAL mode, encrypted credentials via AES-GCM)
- Communication: REST for CRUD, Server-Sent Events for execution progress
- Isolated runs: an environment can send its runs to a throw-away runner (Docker container, AWS ECS Fargate task, Google Cloud Run job execution) that runs the same engine and reports back to the server (15.4)
1.3 Installation
One command per platform. The installer fetches the latest release, verifies a SHA256 checksum, extracts to a user-scope folder, and puts odara on your $PATH. No admin rights needed.
Windows
Paste in PowerShell:
irm https://odara.rs/install.ps1 | iex
Linux
Paste in bash or sh (x86_64). The release is also published as .deb and .rpm packages.
curl -fsSL https://odara.rs/install.sh | sh
macOS
Coming soon. Linux (x86_64) and Windows are available today; macOS and ARM64 are on the roadmap.
After install, open a new terminal, run odara, and visit http://localhost:3002.
Options
Both scripts read these environment variables:
| Variable | Default / Notes |
|---|---|
ODARA_VERSION | latest — set to X.Y.Z to pin a specific release |
ODARA_INSTALL_DIR | Windows: %LOCALAPPDATA%\Programs\Odara · Linux: ~/.local/share/odara |
ODARA_RUN | Set to 1 to launch immediately after install |
ODARA_BIN_DIR | Linux only: where the odara command goes (default ~/.local/bin) |
ODARA_NO_PATH | Linux only: 1 to leave your shell's PATH untouched |
ODARA_NO_LAUNCHER | Linux only: 1 to skip the desktop launcher |
Example — pin a specific version on Windows:
$env:ODARA_VERSION='0.4.1'; irm https://odara.rs/install.ps1 | iex
Example — install to a custom path on Linux:
ODARA_INSTALL_DIR=/opt/odara curl -fsSL https://odara.rs/install.sh | sh
1.4 First pipeline in 5 minutes
- Open Odara → create a new project (or use the default).
- In the Explorer, click + New Pipeline.
- From the Nodes palette, drag a
CSV Sourceonto the canvas. - Drag a
Consoletarget next to it, and connect them. - Click the CSV node → set the File Path.
- Click the green Execute button in the top bar.
You'll see real-time progress in the canvas and the rows printed in the Console Log panel. That's it — you have a running pipeline.
1.5 System requirements & limits
| Requirement | Minimum | Recommended |
|---|---|---|
| OS | Linux x86_64 (Ubuntu 22.04+, Debian 12+, Fedora 36+) · Windows 10 64-bit with PowerShell 5.1+ | Ubuntu 24.04 · Windows 11 |
| RAM | 4 GB | 16 GB |
| Disk | 500 MB + project data | 10 GB free |
| Python (for Python transforms) | 3.9+ | 3.12 |