# Where Does It Run?
Source: https://humanspark.ai/frameworks/where-does-it-run/

*Every AI system has a venue, and the venue determines where your data goes, what breaks, what it costs, and what it can do.*

**When to use:** Before building anything that moves past product. If you're regulated, before choosing what to automate at all.

Venue = *where the rules run* + *where the model runs*

The two parts can live in different places. Whether local is good enough for your work is measurable, so [benchmark it rather than guess](https://humanspark.ai/local-ai/).

Your rules, their model

The common hybrid - a wrapper on your server, a frontier model at the end of an API.

Both parts yours

Full control, running a local model on your own hardware.

> **Data follows the model:** Rules on your own server does not keep your data home. Every request travels to wherever the model lives - if the model sits in someone else's cloud, so does everything you send it.

> **Filter first:** Venue is usually decided on the way to building. If contracts or your regulator say certain data never leaves the building, venue is the first filter - applied before you choose what to automate.

> **Sits beside:** You've met this in miniature: [The Three Tiers of AI Privacy](https://humanspark.ai/frameworks/three-tiers-of-ai-privacy/) - consumer cloud, enterprise cloud, local - is the data view of venue, and the place to start when you're buying rather than building. This card is the decision behind it: where each of the two parts sits, and everything that follows from that. Not [the Three Ways](https://humanspark.ai/frameworks/three-ways-of-working/): the Ways describe your relationship to the system; venue describes where its two parts sit. Either can change while the other stays put.