strakedata

Strake v0.2.5 is live! 🎉 Get the Python module or via our CLI downloader. github.com/strake-data/st… #DataEngineering #AI

"agents that should 'just work on a laptop' shouldn't need to ship a database container." Total agree about embedded databases for agents! Also, there is a need for sandboxed and governed access for agents that strake brings to the table.

Strake v0.2.4 is live! Rust 2024 edition, one GenericSqlConnector unifying all SQL dialects. Hardened secrets, async circuit breakers, and granular schema drift detection. Cleaner architecture, safer code. github.com/strake-data/st…

@levie This post perfectly describes why we think agents need an "ai data layer". They need sandboxes with code execution, pushdown optimization, and isolation. Not a simple query tool but a high performance infrastructure layer where agents actually process data.

It’s wild to think about what types of infrastructure and services must change in a world where agents can process information a hundred or a thousand times faster than humans. Even the tools that were built for machine speed before, generally were still in service of end-users making a request somewhere in the system. Agents running 24/7 and in parallel modify these requirements meaningfully. Here are just a few examples: * Sandboxes. Agents need sandboxes to operate in that have to be insanely low latency because they can boot up these environments for coding at any moment. * Search (both publicly and within an enterprise). Agents can parallelize searches hundreds or thousands of times so they need to be able to work with fast indexes of information. * Payments. Agents can now pay in micro transactions, and aren’t bothered by the friction of paying $0.01 for a resource that a human would be. * File systems. Agents need to be able to work with files at a scale that humans never had to worry about. You’ll have all new complexity around version control, permissions, and having agents reading/writing from data at insane speeds. And there are tons more. We’re going from a word where software was built for people to a world where it’s built for agents. Lots of changes downstream as a result.

MCP is powerful but has a scaling problem: tool definitions and data rows explode your context window. We built Strake around the "code execution pattern". Agents write code to interact with MCP servers, process data locally, return only what matters. strakedata.substack.com/p/stop-sending…

Strake is the AI data layer: sandboxed Python execution on live data. No ETL, just zero-copy. This DevOps agent demo correlates SQL, Parquet, JSON, CSV, and APIs without pipelines. Plans before it acts. Runs code where your data lives. #mcp #agenticai github.com/strake-data/st…

If you're giving LLMs write access to your infrastructure, you're one hallucination away from a disaster. We think the solution is a sandboxed, federated read-only SQL interface. Read our take: strakedata.substack.com/p/stop-sending…

The first blog post on how we see Strake as a solution for token efficient and safe way of using AI agents in production #ai #mcp open.substack.com/pub/strakedata…