> For the complete documentation index, see [llms.txt](https://docs.eyepop.ai/developer-documentation/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.eyepop.ai/developer-documentation/deploying.md).

# Deploying

Choose where your Pops run

A Pop definition is the same wherever it runs. What changes is who operates the machine it runs on.

|               | [Cloud](/developer-documentation/deploying/cloud.md)                       | [On-Premise](/developer-documentation/deploying/on-premise.md)                     |
| ------------- | -------------------------------------------------------------------------- | ---------------------------------------------------------------------------------- |
| Runs on       | infrastructure EyePop hosts                                                | a machine you control                                                              |
| Media goes    | to EyePop                                                                  | nowhere — it stays on the host                                                     |
| You provision | nothing                                                                    | one Linux host with Docker                                                         |
| Scaling       | on demand                                                                  | the hardware you have                                                              |
| Best for      | getting started, bursty and batch work, production without hardware to run | data that cannot leave your network, cameras on a private LAN, latency at the edge |

Most projects start in the cloud. Move on-premise when the media cannot leave your network, or when the hardware is already where the cameras are.

### Cloud

Two session types. A **transient** session, created by the SDK on demand, is the right default while you build. A **Deployment** is persistent, with your Pop's models pre-loaded, for production latency.

* [Cloud](/developer-documentation/deploying/cloud.md) — how the two compare
* [Development](/developer-documentation/deploying/cloud/development.md) — connect the SDK and run a first inference
* [Deployment](/developer-documentation/deploying/cloud/deployment.md) — create a persistent session and attach to it
* [Private Cameras](/developer-documentation/deploying/cloud/private-cameras.md) — cloud inference for cameras behind a firewall

### On-Premise

An **instance** is the EyePop runtime installed on one machine, serving one Pop. The CLI creates and manages it, and the SDKs reach it in local mode.

* [On-Premise](/developer-documentation/deploying/on-premise.md) — create an instance, change its Pop, and run inference against it

### Next steps

* [Pops](/developer-documentation/platform/pop.md) — the definition that runs in either place
* [Pricing](/developer-documentation/pricing.md) — what each one costs