nao

Sounds cool, I want to try this kind of things out, but do you have or planned to have a sandboxing environment, where the agent can try running the query in let say duckdb first to confirm its validity/result before sending it over to bq? Or use something like tablesample when developing the query to reduce cost? One more thing, how do you compare with nao ( https://github.com/getnao/nao ), it's something I've followed for a while and seem to answer similar issue as what ktx build

I am building nao, an open-source analytics agent. With nao you can connect all your sources of context (from warehouse to your knowledge base) using a CLI, and then you have an agent that sits on top of it. You can use the agent from any client (web, Slack, Teams but also other harnesses). We think most of data analytics work will be transformed (and is already being transformed) from SQL monkeys to chat to analyses, but all the UI/UX are not designed for this and this is what nao is, being open-source because knowing how the context is managed is key. With nao you can have a conversation and then shared the output on the form of a story, that can be either static or live replacing what old dashboards were. We are close to 1k stars on Github: https://github.com/getnao/nao

I'm working most of my time on nao [1] (an open-source analytics agent). With nao we are exploring the fact that agents are good to work with filesystems, so we help people getting the context into the filesystem and then you can plug our agent loop on top (on anything else). The most exciting part is working with sandboxes and emulating filesystems. So at the moment the filesystem is local, but it could be a bucket or a database that we emulate as real filesystem for the agentic loop, all of this can also be mounted to sandboxes, and in the end you can do data transformations all in isolation with control on the context. [1] https://github.com/getnao/nao

OpenAI recently shared how their internal data agent works: [https://openai.com/index/inside-our-in-house-data-agent/](https://openai.com/index/inside-our-in-house-data-agent/) What interested me most wasn’t the model itself, but the surrounding system design: automated context retrieval, evaluation loops, and memory that improves the agent over time. I’ve been experimenting with recreating a similar setup, but with a different goal: making the configuration more accessible for any company. The result is an open-source YAML + Markdown framework where you define context sources, tools, and behavior explicitly instead of writing Python code. The idea is to make agent context easier to reason about, version, and iterate on, especially for data teams. Repo: [https://github.com/getnao/nao](https://github.com/getnao/nao) Would love feedback fro