Skip to content

Installation

Strake consists of two main components: 1. Strake Client (Python): For writing queries and interacting with the engine. 2. Strake Server (Rust): The core engine that runs the queries (can also be embedded in the Python client).

Prerequisites

  • Linux or macOS (Windows support is experimental).
  • Python 3.9+
  • Rust 1.75+ (Only required if building from source).

1. Installing the Python Client

Strake provides a high-performance Python client written in Rust.

## uv
uv init strake-demo
uv pip add strake
or 
## pip
pip install strake

2. Universal Install Script (Linux/macOS)

The fastest way to install Strake binaries is via the universal install script:

# Install CLI only (default)
curl -sSfL https://strakedata.com/install.sh | sh

# Install all binaries (CLI, Server, Enterprise)
curl -sSfL https://strakedata.com/install.sh | sh -s -- --all

# Install specific components
curl -sSfL https://strakedata.com/install.sh | sh -s -- --cli --server

# Install a specific version
curl -sSfL https://strakedata.com/install.sh | sh -s -- --cli --version v0.1.0

# Force static (musl) binary on Linux (better for older distros/GCP)
curl -sSfL https://strakedata.com/install.sh | sh -s -- --musl

Options

Flag Description
--all, -a Install all binaries (cli, server, enterprise)
--cli Install strake-cli only
--server Install strake-server only
--enterprise Install strake-enterprise only
--version, -v Install a specific version (e.g., v0.1.0)
--musl Force installation of the statically-linked musl binary (Linux only)

Environment Variables

Variable Default Description
STRAKE_INSTALL_DIR ~/.local/bin Installation directory

Note: Make sure ~/.local/bin is in your PATH. The script will remind you if it's not.

Linux Compatibility: Our standard binaries are built against GLIBC 2.31, making them compatible with most modern and legacy Linux distributions (including Debian 10/11, Ubuntu 20.04+, and common GCP/AWS images). If you encounter GLIBC version errors, use the --musl flag to install a fully static binary.

3. Building from Source

To build the Python bindings locally:

  1. Navigate to the strake-python directory: 
    cd strake-python
  2. Create a virtual environment: 
    python3 -m venv .venv
source .venv/bin/activate
  3. Install build dependencies and build: 
    pip install maturin
maturin develop --release

4. Running the Server

You can run Strake as a standalone server exposing an Apache Arrow Flight SQL interface.

Open Source Edition (OSS)

Run the standard server for local development or basic deployments.

# From the project root
cargo run --package strake-server --release

The server will start on 0.0.0.0:50051.

Server Configuration

The Strake Server can be configured via environment variables or a .env file. These settings override values in config/strake.yaml.

Variable Default Description
STRAKE_SERVER__LISTEN_ADDR 0.0.0.0:50051 The address and port for the gRPC server
STRAKE_SERVER__HEALTH_ADDR 0.0.0.0:8080 The address and port for the health check and API
STRAKE_SERVER__CATALOG strake The default catalog name
STRAKE_SERVER__GLOBAL_CONNECTION_BUDGET 100 Max concurrent connection budget for the server
STRAKE_API_URL http://localhost:8080/api/v1 Public URL for the Strake API
STRAKE_AI_PROVIDER None AI provider for metadata enrichment (gemini, openai)
STRAKE_AI_MODEL None Overrides the AI model used for descriptions
GOOGLE_API_KEY None API key for Gemini provider
OPENAI_API_KEY None API key for OpenAI provider
STRAKE_AUTH__ENABLED false Enable/Disable authentication checks
STRAKE_AUTH__API_KEY (empty) Static API key for auth (required if enabled)
STRAKE_RETRY__MAX_ATTEMPTS 5 Max retries for establishing upstream connections
STRAKE_QUERY_LIMITS__MAX_OUTPUT_ROWS None Hard limit on rows returned by a query

For detailed instructions on setting up database-backed authentication and generating keys, see the Authentication Guide.