> ## Documentation Index > Fetch the complete documentation index at: https://docs.2501.ai/llms.txt > Use this file to discover all available pages before exploring further. # Overview > Install and run 2501 on your own infrastructure from the signed files of each build Every build of 2501, `X.Y.Z-r` (for example `0.19.0-r1`), is published as a Release of [github.com/2501-ai/install](https://github.com/2501-ai/install). The Release holds the files you install that build from, all generated by 2501 from one Helm chart and signed. What is yours, your settings and your secrets, stays with you and never in our files. | You run | Files of the Release | You keep | | - | - | - | | Kubernetes or OpenShift, with `kubectl` or `oc`, Helm or ArgoCD | `2501-.tgz`, the chart, and `values.example.yaml` | your `values.yaml`, in your Git, and your Secrets | | Docker Swarm | `compose.yaml`, `env.example` | your `.env`, your Docker secrets | On Kubernetes and OpenShift you render the chart with your `values.yaml` on your workstation (`helm template`, with the `helm` binary only) and apply the result with `kubectl` or `oc`; nothing of Helm runs in your cluster. A site that runs Helm or ArgoCD applies the chart with them instead. Each file comes with its signature, `.bundle`. `https://github.com/2501-ai/install/releases/latest/download/` always gives the newest build's file; the Release of a build gives that build's, and never changes once published. ## How an installation works **1. Get the files** Download the files of your platform from the Release of the build you install, and check each signature as [Verify Image Signatures](https://docs.2501.ai/0.19/deployment/verify-downloads.md#verify-the-installation-files) shows. **2. Write your settings and create your secrets** Once: your `values.yaml` from `values.example.yaml` on Kubernetes and OpenShift, or your `.env` from `env.example` on Docker Swarm. On Kubernetes and OpenShift, `values.yaml` needs the platform, the registry, Command Center's host name and TLS mode, and whether 2501 provides the database and Elasticsearch: see [values.yaml](https://docs.2501.ai/0.19/deployment/configuration.md). On Docker Swarm, `compose.yaml` sets those, and `.env` needs Command Center's host name and port and the docker group: see [Docker Swarm](https://docs.2501.ai/0.19/deployment/docker-swarm.md#write-your-env). The tenant, administrator and LLM are optional. Then the secrets: their commands are in [Kubernetes and OpenShift](https://docs.2501.ai/0.19/deployment/kubernetes.md#create-the-secrets) and at the top of `compose.yaml`. The services wait until the secrets exist. **3. Apply** Render the chart and `kubectl apply` or `oc apply` the whole result, or `helm upgrade --install`, or `docker stack deploy`: see [Kubernetes and OpenShift](https://docs.2501.ai/0.19/deployment/kubernetes.md) and [Docker Swarm](https://docs.2501.ai/0.19/deployment/docker-swarm.md). **4. Check and sign in** Wait for the services to be ready, then sign in to Command Center as the first administrator. The first run creates your tenant, organizations and administrator from your settings, or `Default` and `admin@2501.ai` without them, and your LLM model if you gave one; otherwise add the LLM in Command Center. See [Users & Organizations](https://docs.2501.ai/0.19/deployment/users-organizations.md). An upgrade applies the files of a newer build with the same `values.yaml` or `.env`, and keeps your secrets as they are. See [Upgrade](https://docs.2501.ai/0.19/deployment/upgrade.md). ## The images The files name each image by its digest, so a build always runs the images it was tested with. They pull them from 2501's registry, with the access your 2501 account team grants, or from your copy: see [Pull the images from 2501's registry](https://docs.2501.ai/0.19/deployment/pull-from-2501.md). ## Your site What your site needs beyond the standard installation, such as your own database, a corporate CA, a proxy, a registry of yours or another storage class, is a value in your `values.yaml` on Kubernetes and OpenShift: [Your site's values](https://docs.2501.ai/0.19/deployment/kubernetes.md#your-sites-values). On Docker Swarm, you add it in a second compose file of yours: [Adapt to your site](https://docs.2501.ai/0.19/deployment/docker-swarm.md#adapt-to-your-site). 2501 never keeps your values or renders a file for you; your 2501 SA helps you fill them. Windows targets reached through a gMSA (Kerberos) need Docker Swarm: see [gMSA hosts](https://docs.2501.ai/0.19/deployment/docker-swarm.md#gmsa-hosts). ## What you need | | | | - | - | | A place to run | A Kubernetes or OpenShift namespace, or a Docker Swarm. | | A workstation | `kubectl` or `oc`, and the `helm` binary to render the chart, on Kubernetes and OpenShift; `docker` on a Swarm manager. | | Access to the images | 2501's registry, with the access your 2501 account team grants, or a copy in your own registry. | | A database | The PostgreSQL 2501 provides on Kubernetes and Docker Swarm, or your own. OpenShift needs your own. | | An address | The host name users reach Command Center on, with its TLS certificate. | | An LLM | Its URL, model and API key, now or later in Command Center. | ## Support A version is supported for 3 months after its release: each Release says until when, under "Supported until". Fixes go into the newest version only, so a site past that date upgrades first. ## Coming from the 2501 CLI The `2501 infra` CLI no longer installs or upgrades 2501. An installation it created moves to the chart once, in place: the chart updates its objects, keeping their data and its keys. See [Move from the CLI](https://docs.2501.ai/0.19/deployment/migrate-from-cli.md). > [card] [Kubernetes and OpenShift](https://docs.2501.ai/0.19/deployment/kubernetes.md) > > Render the chart with your values.yaml and apply it, or use Helm or ArgoCD > [card] [Docker Swarm](https://docs.2501.ai/0.19/deployment/docker-swarm.md) > > Fill your .env and deploy the stack > [card] [values.yaml](https://docs.2501.ai/0.19/deployment/configuration.md) > > Every value and secret of an installation > [card] [Upgrade](https://docs.2501.ai/0.19/deployment/upgrade.md) > > Read, review, apply, roll back > [card] [Operations](https://docs.2501.ai/0.19/deployment/operations.md) > > Feature toggles, webhooks, sign-in recovery, license > [card] [Troubleshooting](https://docs.2501.ai/0.19/deployment/troubleshooting.md) > > Common issues and solutions This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.