diff --git a/server/README.md b/server/README.md
index 40b479577..66e17e5bd 100644
--- a/server/README.md
+++ b/server/README.md
@@ -38,7 +38,13 @@ And ping again
 This time you'll see the updated message.
 
 For more details about how to get museum up and running, see
-[RUNNING.md](RUNNING.md).
+[RUNNING](RUNNING.md).
+
+> [!TIP]
+>
+> Also, there is a way to use our pre-built Docker images to directly start a
+> cluster without needing to clone this repository - see
+> [docs/docker](docs/docker.md).
 
 ## Architecture
 
@@ -84,10 +90,11 @@ And it is built with containerization in mind - both during development and
 deployment. Just use the provided Dockerfile, configure to taste and you're off
 to the races.
 
-> [!CAUTION]
->
-> We don't publish any official docker images (yet). For self-hosters, the
-> recommendation is to build your own image using the provided `Dockerfile`.
+Overall, there are [three approaches](RUNNING.md) you can take:
+
+* Run using Docker using a pre-built Docker image
+* Run using Docker but build an image from source
+* Run without Docker
 
 Everything that you might needed to run museum is all in here, since this is the
 setup we ourselves use in production.
diff --git a/server/RUNNING.md b/server/RUNNING.md
index 22045fe2b..132bc7801 100644
--- a/server/RUNNING.md
+++ b/server/RUNNING.md
@@ -8,13 +8,14 @@ environment that doesn't clutter your machine.
 You can also run museum directly on your machine if you wish - it is a single
 static go binary.
 
-This document describes both these approaches, and also outlines configuration.
+This document describes these approaches, and also outlines configuration.
 
-- [Running using Docker](#docker)
-- [Running without Docker](#without-docker)
+- [Run using Docker using a pre-built Docker image](docs/docker.md)
+- [Run using Docker but build an image from source](#build-and-run-using-docker)
+- [Running without Docker](#run-without-docker)
 - [Configuration](#configuration)
 
-## Docker
+## Build and run using Docker
 
 Start the cluster
 
@@ -70,7 +71,7 @@ Each time museum gets rebuilt from source, a new image gets created but the old
 one is retained as a dangling image. You can use `docker image prune --force`,
 or `docker system prune` if that's fine with you, to remove these.
 
-## Without Docker
+## Running without Docker
 
 The museum binary can be run by using `go run cmd/museum/main.go`. But first,
 you'll need to prepare your machine for development. Here we give the steps,
@@ -132,7 +133,7 @@ pg_ctl -D /usr/local/var/postgres -l logfile start
 createuser -s postgres
 ```
 
-## Start museum
+### Start museum
 
 ```sh
 export ENTE_DB_USER=postgres
@@ -148,7 +149,7 @@ ENTE_DB_USER=ente_user
 air
 ```
 
-## Testing
+### Testing
 
 Set up a local database for testing. This is not required for running the server.
 Create a test database with the following name and credentials:
diff --git a/server/docs/docker.md b/server/docs/docker.md
index c1c72ea20..d8f3db913 100644
--- a/server/docs/docker.md
+++ b/server/docs/docker.md
@@ -52,7 +52,7 @@ require you to clone the repository or build any images.
    touch museum.yaml
    ```
 
-4. That is all. You can now start everything.
+5. That is all. You can now start everything.
 
    ```sh
    docker compose up
@@ -66,9 +66,18 @@ This will start a cluster containing:
 
 For each of these, it'll use the latest published Docker image.
 
-Alternatively, if you only want to run Ente's server, you can directly pull and
-run the image we publish to **`ghcr.io/ente-io/server`**.
+You can do a quick smoke test by pinging the API:
 
 ```sh
-docker pull ghcr.io/ente-io/server:latest
+curl localhost:8080/ping
+```
+
+## Only the server
+
+Alternatively, if you have setup the database and object storage externally and
+only want to run Ente's server, you can skip the steps above and directly pull
+and run the image from **`ghcr.io/ente-io/server`**.
+
+```sh
+docker pull ghcr.io/ente-io/server
 ```