strake 0.2.2

The AI Data Layer: Secure, sandboxed environment for AI agents to query and process data.

Strake

The AI Data Layer

Strake is the AI Data Layer. Not just a query tool, and not a RAG pipeline. It's the sandboxed execution environment where agents meet your data and return answers, not rows.

Built on Apache Arrow DataFusion, Strake enables AI agents to discover, query, and process data across your entire stack (PostgreSQL, Snowflake, S3, and more) without the need for data movement or ETL.

📚 Full Documentation: Check out the complete documentation for installation, architecture, and API references.

---

Key Features

• Developer First: Built for engineers. Type-safe configuration, rich CLI tooling, and local development workflows.
• Secure Execution Layer: Run untrusted Python code safely using Firecracker MicroVMs or Native OS Sandboxing (Landlock, Seccomp, Namespaces).
• High Performance: Sub-second latency for federated joins using Apache Arrow.
• Pluggable Sources: Postgres, S3, Local Files, REST, gRPC, and more.
• MCP-Native Discovery: Built for the Model Context Protocol. Your agents discover your entire data catalog and schemas instantly.
• Python Native: Zero-copy integration with Pandas and Polars via PyO3.
• Enterprise Governance: Row-Level Security (RLS), Column Masking, and OIDC Authentication (Enterprise Edition).
• Observability: Built-in OpenTelemetry tracing and Prometheus metrics.
• GitOps Native: Manage your data mesh configuration as code. Version control your sources, policies, and metrics.
• Enterprise Features: OIDC, Row-Level Security, and Data Contracts (see Enterprise Edition).

Code Mode: Don't Compute in Context

Most agents fail by swallowing thousands of raw SQL rows. Strake's Code Mode lets them process data in Python inside a secure sandbox, sending only the parsed results to the LLM.

import strake
from strake.mcp import run_python

# Query 10M rows instantly via DataFusion
# Aggregate in Python to prevent context bloat
script = """
df = strake.sql("SELECT * FROM user_events")
summary = df.groupby('feature_flag')['latency'].median()
print(summary.to_json())
"""

# Runs isolated with OS Sandboxing or Firecracker VMs
result = await run_python(script)

Quick Start (5-Minute Setup)

1. Installation

Quick Install (Linux/macOS)

curl -sSfL https://strakedata.com/install.sh | sh

Install via Cargo (Rust)

cargo install --path crates/cli
cargo install --path crates/server

Python Client

pip install strake

2. Configuration (GitOps)

Initialize a new config and validate your sources:

# Initialize a new config
strake-cli init

# Validate configuration
strake-cli validate sources.yaml

# Apply to the metadata store (Sync)
strake-cli apply sources.yaml --force

3. Query with Python

First, define your data sources in a sources.yaml file:

sources:
  - name: local_files
    type: csv
    path: "data/*.csv"
    has_header: true
    tables:
      - name: measurements

Then, query using the Strake Python client:

import strake
import polars as pl

# Connect using your source configuration
conn = strake.connect(sources_config="sources.yaml")

# Query across sources using standard SQL
query = "SELECT * FROM measurements LIMIT 5"
data = conn.sql(query)

# Zero-copy integration with Polars/Pandas
df = pl.from_arrow(data)
print(df)

Project Structure

Component	Description
strake-runtime	Orchestration layer (Federation Engine, Sidecar).
strake-connectors	Data source implementations (Postgres, S3, REST, etc).
strake-sql	SQL Dialects, Query Optimization, and Substrait generation.
strake-common	Shared types, configuration, and error handling.
strake-server	Arrow Flight SQL server implementation.
strake-cli	GitOps CLI for managing data mesh configurations.
strake-python	Python bindings for high-performance data access.

Contributing

We welcome contributions! Please see our Contributing Guidelines for details on how to get started.

License

Strake is licensed under the Apache 2.0 license.