diff --git a/docs/cli/Guides/deploy-app/deploy-app-example.md b/docs/cli/Guides/deploy-app/deploy-app-example.md
index 762949db..edb21fd1 100644
--- a/docs/cli/Guides/deploy-app/deploy-app-example.md
+++ b/docs/cli/Guides/deploy-app/deploy-app-example.md
@@ -98,7 +98,7 @@ RUN chmod +x /usr/local/bin/usd_to_crypto.py
 COPY entrypoint.sh /usr/local/bin/entrypoint.sh
 RUN chmod +x /usr/local/bin/entrypoint.sh
 
-# Set /sp as workdir (doesn't matter in this case; entrypoint.sh uses /sp/output as workdir)
+# Set /sp as workdir (doesn't matter in this case -- entrypoint.sh uses /sp/output as workdir)
 WORKDIR /sp
 
 # Set entrypoint
@@ -137,8 +137,8 @@ tar -czvf input.tar.gz ./input.txt
 
 ```shell
 ./spctl files upload ./input.tar.gz \
---filename input.tar.gz \
---output input.resource.json
+    --filename input.tar.gz \
+    --output input.resource.json
 ```
 
 ### 3. Deploy
@@ -147,9 +147,9 @@ Place an order:
 
 ```shell
 ./spctl workflows create \
---tee 7 \
---solution ./usd_to_crypto.resource.json \
---data ./input.resource.json
+    --tee 7 \
+    --solution ./usd_to_crypto.resource.json \
+    --data ./input.resource.json
 ```
 
 Find the order ID in the output, for example:
diff --git a/docs/cli/Guides/swarm-vllm.md b/docs/cli/Guides/swarm-vllm.md
new file mode 100644
index 00000000..52f4a9ae
--- /dev/null
+++ b/docs/cli/Guides/swarm-vllm.md
@@ -0,0 +1,183 @@
+---
+id: "swarm-vllm"
+title: "vLLM on Super Swarm"
+slug: "/guides/swarm-vllm"
+sidebar_position: 20
+---
+
+This guide provides step-by-step instructions for deploying MedGemma and Apertus on Super Swarm using vLLM.
+
+## Prerequisites
+
+- [kubectl](https://kubernetes.io/docs/tasks/tools/)
+- [helm](https://helm.sh/docs/intro/install/)
+- A domain
+- For [MedGemma](https://huggingface.co/google/medgemma-1.5-4b-it): a Hugging Face token from an account that has already accepted the model's terms
+
+Also, download and rename deployment scripts:
+
+- [`deploy_medgemma_official.sh`](/files/deploy_medgemma_official.sh)
+- [`deploy_apertus_official.sh`](/files/deploy_apertus_official.sh)
+
+## 1. Sign in to Super Swarm
+
+In the Super Swarm dashboard, sign in using MetaMask:
+
+<img src={require('../images/swarm-log-in.png').default} width="auto" height="auto" border="1"/>
+<br/>
+
+## 2. Create a Kubernetes cluster
+
+2.1. Go to **Kubernetes** and press **Create Cluster**:
+
+<img src={require('../images/kubernetes-create-cluster.png').default} width="auto" height="auto" border="1"/>
+<br/>
+<br/>
+
+2.2. Add a GPU to the cluster, allocate resources, and press **Create Cluster**:
+
+<img src={require('../images/create-kubernetes-space.png').default} width="auto" height="auto" border="1"/>
+<br/>
+
+## 3. Download the cluster configuration file
+
+<img src={require('../images/kubernetes-download-kubeconfig.png').default} width="auto" height="auto" border="1"/>
+<br/>
+
+## 4. Point `kubectl` to the configuration file
+
+Execute the following command:
+
+```shell
+export KUBECONFIG=<xxxx-xxx-xxx-xxxx>-kubeconfig.yaml
+```
+
+Replace `<xxxx-xxx-xxx-xxxx>-kubeconfig.yaml` with the name of the downloaded configuration file.
+
+## 5. Update the scripts
+
+In both scripts (`deploy_medgemma_official.sh` and `deploy_apertus_official.sh`), find `BASE_DOMAIN="${BASE_DOMAIN:-monai-swarm.win}"` and replace `monai-swarm.win` with your domain.
+
+## 6. Set the API key
+
+Choose any password that will protect your API endpoints. Execute the following command and type your chosen secret (characters won't be displayed):
+
+```shell
+read -rs API_KEY && export API_KEY
+```
+
+## 7. Deploy the model
+
+### Apertus
+
+```shell
+bash deploy_apertus_official.sh
+```
+
+The deployment usually takes 5-7 minutes.
+
+A working Apertus config is already set in the script:
+
+```
+dtype=bfloat16
+max-model-len=32768
+gpu-memory-utilization=0.55
+max-num-seqs=8
+max-num-batched-tokens=4096
+```
+
+### MedGemma
+
+```shell
+export HF_TOKEN=hf_xxx
+bash deploy_medgemma_official.sh
+```
+
+Replace `hf_xxx` with an HF_TOKEN.
+
+Alternatively, create a `.hf_token` file with the token next to `deploy_medgemma_official.sh`; the script will read it automatically.
+
+A working MedGemma config is already set in the script:
+
+```
+dtype=bfloat16
+max-model-len=8192
+gpu-memory-utilization=0.40
+--mm-processor-cache-gb 1
+max-num-seqs=4
+max-num-batched-tokens=2048
+```
+
+## 8. Check Kubernetes
+
+```shell
+kubectl get pods -o wide
+kubectl get svc
+kubectl get ingress
+```
+
+Expected output:
+
+- Two pods in `1/1 Running`
+- Two services
+- Two ingresses
+
+## 9. Confirm DNS records
+
+Back in the Super Swarm dashboard, go to **Ingresses** and note the two hostnames listed there.
+
+<img src={require('../images/ingresses.png').default} width="auto" height="auto" border="1"/>
+<br/>
+<br/>
+
+For each hostname, add a CNAME record pointing to it and a TXT record for domain verification at your DNS provider.
+
+## 10. Publish the cluster
+
+In the Super Swarm dashboard, go to **Kubernetes** and publish the cluster.
+
+<img src={require('../images/kubernetes-publish-cluster.png').default} width="auto" height="auto" border="1"/>
+<br/>
+
+## 11. Send test requests
+
+In the test requests below, replace:
+
+- `<DOMAIN>` with your domain.
+- `<API_KEY>` with the key you set in [Step 6](/cli/guides/swarm-vllm#6-set-the-api-key).
+
+### Apertus
+
+```shell
+curl https://apertus-vllm.<DOMAIN>/v1/completions \
+  -H 'Authorization: Bearer <API_KEY>' \
+  -H 'Content-Type: application/json' \
+  -d '{
+    "model": "swiss-ai/Apertus-8B-2509",
+    "prompt": "Write a concise technical summary of Kubernetes GPU scheduling.",
+    "temperature": 0,
+    "max_tokens": 200
+  }'
+```
+
+### MedGemma
+
+```shell
+curl https://medgemma-vllm.<DOMAIN>/v1/chat/completions \
+  -H 'Authorization: Bearer <API_KEY>' \
+  -H 'Content-Type: application/json' \
+  -d '{
+    "model": "google/medgemma-1.5-4b-it",
+    "messages": [
+      {
+        "role": "user",
+        "content": [
+          {"type": "text", "text": "Describe this image briefly."},
+          {"type": "image_url", "image_url": {"url": "data:image/png;base64,PASTE_BASE64_HERE"}} 
+        ]
+      }
+    ],
+    "temperature": 0,
+    "max_tokens": 120
+  }'
+```
\ No newline at end of file
diff --git a/docs/cli/images/create-kubernetes-space.png b/docs/cli/images/create-kubernetes-space.png
new file mode 100644
index 00000000..c98291d6
Binary files /dev/null and b/docs/cli/images/create-kubernetes-space.png differ
diff --git a/docs/cli/images/ingresses.png b/docs/cli/images/ingresses.png
new file mode 100644
index 00000000..86b2a44b
Binary files /dev/null and b/docs/cli/images/ingresses.png differ
diff --git a/docs/cli/images/kubernetes-create-cluster.png b/docs/cli/images/kubernetes-create-cluster.png
new file mode 100644
index 00000000..312d03da
Binary files /dev/null and b/docs/cli/images/kubernetes-create-cluster.png differ
diff --git a/docs/cli/images/kubernetes-download-kubeconfig.png b/docs/cli/images/kubernetes-download-kubeconfig.png
new file mode 100644
index 00000000..cf287367
Binary files /dev/null and b/docs/cli/images/kubernetes-download-kubeconfig.png differ
diff --git a/docs/cli/images/kubernetes-publish-cluster.png b/docs/cli/images/kubernetes-publish-cluster.png
new file mode 100644
index 00000000..bf0c2bfb
Binary files /dev/null and b/docs/cli/images/kubernetes-publish-cluster.png differ
diff --git a/docs/cli/images/swarm-log-in.png b/docs/cli/images/swarm-log-in.png
new file mode 100644
index 00000000..e7abee2f
Binary files /dev/null and b/docs/cli/images/swarm-log-in.png differ
diff --git a/docs/marketplace/Guides/storage.md b/docs/marketplace/Guides/storage.md
index 95915151..21ebb997 100644
--- a/docs/marketplace/Guides/storage.md
+++ b/docs/marketplace/Guides/storage.md
@@ -7,14 +7,12 @@ sidebar_position: 6
 
 This guide provides step-by-step instructions on how to set up your personal Storj account.
 
-The guide is intended for advanced Web3 users; feel free to skip it and continue using the default recommended option—**Super Protocol cloud**. Read about [types of storage](/marketplace/account/web3#storage).
+The guide is intended for advanced users; feel free to skip it and continue using the default recommended option—**Super Protocol cloud**. Read about [types of storage](/marketplace/account#storage).
 
-<img src={require('../images/web3-account.png').default} width="auto" height="auto" border="1"/>
+<img src={require('../images/web2-account.png').default} width="auto" height="auto" border="1"/>
 <br/>
 <br/>
 
-Web2 users must first [log in as a Web3 user](/marketplace/guides/log-in) to be able to upload to a personal Storj account instead of the Super Protocol cloud.
-
 ## Step 1. Register a Storj account
 
 If you don't already have a [Storj](https://www.storj.io/) account, register one. Both free Trial and Pro accounts are suitable. Note that with a Trial account, your files will become inaccessible once the trial period ends.
@@ -36,7 +34,7 @@ As a result, you should have two pairs Access Key + Secret Key.
 
 ## Step 4. Set up your Super Protocol Web3 account
 
-Open the [Marketplace web app](https://marketplace.superprotocol.com/). Log in as a Web3 user and open the **Account** window.
+Open the [Marketplace web app](https://marketplace.superprotocol.com/), sign in and open the **Account** window.
 
 <img src={require('../images/marketplace-mm-account-menu.png').default} width="auto" height="auto" border="1"/>
 <br/>
diff --git a/docs/marketplace/account.md b/docs/marketplace/account.md
new file mode 100644
index 00000000..99f147c1
--- /dev/null
+++ b/docs/marketplace/account.md
@@ -0,0 +1,52 @@
+---
+id: "account"
+title: "Enter Marketplace"
+slug: "/account"
+sidebar_position: 3
+---
+
+Super Protocol supports two login methods:
+
+- Web2 requires an account on one of the supported platforms:
+    - Google
+    - Hugging Face
+    - GitHub
+    - Microsoft
+- Web3 requires a software wallet installed as a browser extension:
+    - MetaMask
+    - Trust Wallet
+
+For instructions on how to set up software wallets and connect them to the Marketplace, read [How to Log In as a Web3 User](/marketplace/guides/log-in).
+
+<img src={require('./images/enter-marketplace.png').default} width="auto" height="auto" border="1"/>
+<br/>
+<br/>
+
+## Account window
+
+This window shows your user account settings.
+
+<img src={require('./images/web2-account.png').default} width="auto" height="auto" border="1"/>
+<br/>
+<br/>
+
+**User ID**: your unique user ID.
+
+**Login**: the OAuth2 provider and your login email address.
+
+The **Get SPPI** button allows you to get tokens necessary to place <a id="order"><span className="dashed-underline">orders</span></a>.
+
+### Storage
+
+You have two options of decentralized storage to upload files:
+
+- **Super Protocol cloud**:
+    - Recommended for most users.
+    - Does not require additional setup.
+    - Uses Super Protocol's Storj account and thus relies on Super Protocol as the storage provider.
+- **Your Storj account**:
+    - Intended for advanced users.
+    - Requires creating and [setting up a Storj account](/marketplace/guides/storage).
+    - Gives sole control over the uploaded content and storage account.
+
+Read [How to Set Up Storage](/marketplace/guides/storage) for step-by-step instructions.
\ No newline at end of file
diff --git a/docs/marketplace/account/index.md b/docs/marketplace/account/index.md
deleted file mode 100644
index e2ae472f..00000000
--- a/docs/marketplace/account/index.md
+++ /dev/null
@@ -1,39 +0,0 @@
----
-id: "account"
-title: "Enter Marketplace"
-slug: "/account"
-sidebar_position: 3
----
-
-There are two types of accounts in Super Protocol:
-
-- Web3 User account
-- Web2 User account
-
-<img src={require('../images/enter-marketplace.png').default} width="auto" height="auto" border="1"/>
-<br/>
-<br/>
-
-## Web3 User account
-
-_Web3 User account_ provides access to all Marketplace capabilities, including:
-
-- Full decentralization and sole control of user's funds, models, and datasets.
-- Ability to upload models and datasets to the Super Protocol cloud or a personal Storj account.
-- Placement of <a id="order"><span className="dashed-underline">orders</span></a> using Marketplace <a id="offer"><span className="dashed-underline">offers</span></a> or the user's own uploaded content.
-- Registration of individual providers.
-- Creation and monetization of model and dataset offers on the Marketplace.
-- Ability to request additional SPPI tokens.
-
-Read [How to Log In as a Web3 User](/marketplace/guides/log-in) for step-by-step instructions.
-
-## Web2 User account
-
-_Web2 User account_ is a quick way to start with the Marketplace. It streamlines a few steps, but this comes at the expense of full decentralization, such as using OAuth2 authentication for login instead of the decentralized MetaMask.
-
-To log in as a Web2 user, you need an account on one of the supported platforms:
-
-- Google
-- Hugging Face
-- GitHub
-- Microsoft
diff --git a/docs/marketplace/account/web2.md b/docs/marketplace/account/web2.md
deleted file mode 100644
index 5e2e2140..00000000
--- a/docs/marketplace/account/web2.md
+++ /dev/null
@@ -1,31 +0,0 @@
----
-id: "web2"
-title: "Web2 User Account"
-slug: "/account/web2"
-sidebar_position: 2
----
-
-This window shows the settings of your [Web2 User account](/marketplace/account#web2-user-account).
-
-<img src={require('../images/web2-account.png').default} width="auto" height="auto" border="1"/>
-<br/>
-<br/>
-
-**User ID**: your unique user ID.
-
-**Login**: the OAuth2 provider and your login email address.
-
-## Storage
-
-Super Protocol supports two options of decentralized storage to upload files:
-
-- **Super Protocol cloud**:
-    - Does not require additional setup.
-    - Uses Super Protocol's Storj account and thus relies on Super Protocol as the storage provider.
-    - Costs SPPI tokens for additional storage beyond the basic free package.
-- **Your Storj account**:
-    - Available to Web3 users only.
-    - Requires creating and setting up a Storj account.
-    - Gives sole control over the uploaded content and storage account.
-
-To enable uploading to your personal Storj account, [log in as a Web3 user](/marketplace/guides/log-in).
\ No newline at end of file
diff --git a/docs/marketplace/account/web3.md b/docs/marketplace/account/web3.md
deleted file mode 100644
index dd18bbb3..00000000
--- a/docs/marketplace/account/web3.md
+++ /dev/null
@@ -1,36 +0,0 @@
----
-id: "web3"
-title: "Web3 User Account"
-slug: "/account/web3"
-sidebar_position: 1
----
-
-This window allows you to manage your [Web3 User account](/marketplace/account#web3-user-account).
-
-<img src={require('../images/web3-account.png').default} width="auto" height="auto" border="1"/>
-<br/>
-<br/>
-
-**User ID**: your unique user ID is your EVM wallet address.
-
-**Login**: the Web3 login method and the EVM wallet address you are using. Currently, Super Protocol only supports MetaMask as a Web3 login method.
-
-**Get SPPI** and **Get BNB** buttons allow you to get tokens necessary to place <a id="order"><span className="dashed-underline">orders</span></a>:
-
-- SPPI tokens are required to pay and receive payments in Super Protocol.
-- BNB tokens are required to pay for opBNB blockchain transactions.
-
-## Storage
-
-You have two options of decentralized storage to upload files:
-
-- **Super Protocol cloud**:
-    - Recommended for most users.
-    - Does not require additional setup.
-    - Uses Super Protocol's Storj account and thus relies on Super Protocol as the storage provider.
-- **Your Storj account**:
-    - Intended for advanced users.
-    - Requires creating and setting up a Storj account.
-    - Gives sole control over the uploaded content and storage account.
-
-Read [How to Set Up Storage](/marketplace/guides/storage) for step-by-step instructions.
\ No newline at end of file
diff --git a/docs/marketplace/images/web2-account.png b/docs/marketplace/images/web2-account.png
index 528db2d7..80cc8bee 100644
Binary files a/docs/marketplace/images/web2-account.png and b/docs/marketplace/images/web2-account.png differ
diff --git a/docs/marketplace/images/web3-account.png b/docs/marketplace/images/web3-account.png
deleted file mode 100644
index 46480abe..00000000
Binary files a/docs/marketplace/images/web3-account.png and /dev/null differ
diff --git a/docusaurus.config.js b/docusaurus.config.js
index db773ef3..e38429ec 100644
--- a/docusaurus.config.js
+++ b/docusaurus.config.js
@@ -46,6 +46,14 @@ const config = {
             from: "/cli/guides/quick-guide",
             to: "/cli/guides/deploy-app",
           },
+          {
+            from: "/marketplace/account/web2",
+            to: "/marketplace/account#account-window"
+          },
+          {
+            from: "/marketplace/account/web3",
+            to: "/marketplace/account#account-window"
+          },
         ],
       },
     ],
diff --git a/static/files/deploy_apertus_official.sh b/static/files/deploy_apertus_official.sh
new file mode 100755
index 00000000..1487a1c7
--- /dev/null
+++ b/static/files/deploy_apertus_official.sh
@@ -0,0 +1,154 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+
+BASE_DOMAIN="${BASE_DOMAIN:-monai-swarm.win}"
+API_HOST="${API_HOST:-apertus-vllm.${BASE_DOMAIN}}"
+MODEL_NAME="${MODEL_NAME:-swiss-ai/Apertus-8B-2509}"
+MODEL_ENTRY_NAME="${MODEL_ENTRY_NAME:-apertus}"
+RELEASE_NAME="${RELEASE_NAME:-apertus-official}"
+if [ -z "${API_KEY:-}" ]; then
+  echo "API_KEY must be set. Execute:" >&2
+  echo "read -rs API_KEY && export API_KEY" >&2
+  echo "And then type a desired key." >&2
+  exit 1
+fi
+IMAGE_REPOSITORY="${IMAGE_REPOSITORY:-vllm/vllm-openai}"
+IMAGE_TAG="${IMAGE_TAG:-v0.18.0}"
+GPU_MEMORY_UTILIZATION="${GPU_MEMORY_UTILIZATION:-0.55}"
+MAX_MODEL_LEN="${MAX_MODEL_LEN:-32768}"
+CPU_REQUEST="${CPU_REQUEST:-8}"
+MEMORY_REQUEST="${MEMORY_REQUEST:-48Gi}"
+GPU_COUNT="${GPU_COUNT:-1}"
+PVC_STORAGE="${PVC_STORAGE:-80Gi}"
+INGRESS_CLASS="${INGRESS_CLASS:-nginx}"
+
+need() { command -v "$1" >/dev/null 2>&1 || { echo "Missing dependency: $1" >&2; exit 1; }; }
+need kubectl
+need helm
+
+NAMESPACE="${NAMESPACE:-$(kubectl config view --minify -o jsonpath='{..namespace}' 2>/dev/null || true)}"
+if [ -z "${NAMESPACE}" ]; then
+  NAMESPACE="llm"
+fi
+
+SECRET_NAME="${RELEASE_NAME}-auth"
+SERVICE_NAME="${RELEASE_NAME}-${MODEL_ENTRY_NAME}-engine-service"
+DEPLOY_LABEL_MODEL="${MODEL_ENTRY_NAME}"
+INGRESS_NAME="${RELEASE_NAME}-api-ingress"
+
+echo "==> Runtime: vLLM (official helm chart)"
+echo "==> Namespace: ${NAMESPACE}"
+echo "==> Release: ${RELEASE_NAME}"
+echo "==> API host: ${API_HOST}"
+echo "==> Model: ${MODEL_NAME}"
+echo "==> Model entry name: ${MODEL_ENTRY_NAME}"
+echo "==> Image: ${IMAGE_REPOSITORY}:${IMAGE_TAG}"
+echo "==> Max model length: ${MAX_MODEL_LEN}"
+echo "==> GPU memory utilization: ${GPU_MEMORY_UTILIZATION}"
+echo
+
+kubectl get ns "${NAMESPACE}" >/dev/null 2>&1 || kubectl create ns "${NAMESPACE}"
+
+helm repo add vllm https://vllm-project.github.io/production-stack >/dev/null 2>&1 || true
+helm repo update >/dev/null 2>&1
+
+cat <<EOF | kubectl -n "${NAMESPACE}" apply -f -
+apiVersion: v1
+kind: Secret
+metadata:
+  name: ${SECRET_NAME}
+type: Opaque
+stringData:
+  VLLM_API_KEY: ${API_KEY}
+EOF
+
+VALUES_FILE="$(mktemp)"
+cleanup() {
+  rm -f "${VALUES_FILE}"
+}
+trap cleanup EXIT
+
+cat > "${VALUES_FILE}" <<EOF
+servingEngineSpec:
+  runtimeClassName: nvidia
+  vllmApiKey:
+    secretName: ${SECRET_NAME}
+    secretKey: VLLM_API_KEY
+  modelSpec:
+    - name: ${MODEL_ENTRY_NAME}
+      repository: ${IMAGE_REPOSITORY}
+      tag: ${IMAGE_TAG}
+      modelURL: ${MODEL_NAME}
+      replicaCount: 1
+      requestCPU: ${CPU_REQUEST}
+      requestMemory: "${MEMORY_REQUEST}"
+      requestGPU: ${GPU_COUNT}
+      pvcStorage: "${PVC_STORAGE}"
+      vllmConfig:
+        dtype: bfloat16
+        maxModelLen: ${MAX_MODEL_LEN}
+        gpuMemoryUtilization: ${GPU_MEMORY_UTILIZATION}
+        enablePrefixCaching: true
+        enableChunkedPrefill: true
+        maxNumSeqs: 8
+        extraArgs:
+          - --max-num-batched-tokens
+          - "4096"
+routerSpec:
+  enableRouter: false
+EOF
+
+KUBECONFIG="${KUBECONFIG:-}" helm upgrade --install "${RELEASE_NAME}" vllm/vllm-stack \
+  --namespace "${NAMESPACE}" \
+  -f "${VALUES_FILE}" \
+  --skip-crds \
+  --wait --timeout=20m
+
+cat <<EOF | kubectl -n "${NAMESPACE}" apply -f -
+apiVersion: networking.k8s.io/v1
+kind: Ingress
+metadata:
+  name: ${INGRESS_NAME}
+  annotations:
+    kubernetes.io/ingress.class: ${INGRESS_CLASS}
+    nginx.ingress.kubernetes.io/proxy-body-size: "0"
+    nginx.ingress.kubernetes.io/proxy-read-timeout: "600"
+    nginx.ingress.kubernetes.io/proxy-send-timeout: "600"
+    nginx.ingress.kubernetes.io/proxy-connect-timeout: "600"
+    nginx.ingress.kubernetes.io/proxy-buffering: "off"
+spec:
+  rules:
+  - host: ${API_HOST}
+    http:
+      paths:
+      - path: /
+        pathType: Prefix
+        backend:
+          service:
+            name: ${SERVICE_NAME}
+            port:
+              number: 80
+EOF
+
+echo
+echo "==> Pods:"
+kubectl -n "${NAMESPACE}" get pods -o wide
+echo
+echo "==> Services:"
+kubectl -n "${NAMESPACE}" get svc -o wide
+echo
+echo "==> Ingress:"
+kubectl -n "${NAMESPACE}" get ingress -o wide
+echo
+echo "==> Waiting for Apertus pod readiness..."
+kubectl -n "${NAMESPACE}" wait --for=condition=ready pod \
+  -l "model=${DEPLOY_LABEL_MODEL},helm-release-name=${RELEASE_NAME}" \
+  --timeout=900s
+echo
+echo "==> Ready"
+echo "Base URL: https://${API_HOST}/v1"
+echo "Model: ${MODEL_NAME}"
+echo "Example:"
+echo "  curl https://${API_HOST}/v1/models -H 'Authorization: Bearer ${API_KEY}'"
diff --git a/static/files/deploy_medgemma_official.sh b/static/files/deploy_medgemma_official.sh
new file mode 100755
index 00000000..7845a04e
--- /dev/null
+++ b/static/files/deploy_medgemma_official.sh
@@ -0,0 +1,170 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+
+BASE_DOMAIN="${BASE_DOMAIN:-monai-swarm.win}"
+API_HOST="${API_HOST:-medgemma-vllm.${BASE_DOMAIN}}"
+MODEL_NAME="${MODEL_NAME:-google/medgemma-1.5-4b-it}"
+MODEL_ENTRY_NAME="${MODEL_ENTRY_NAME:-medgemma}"
+RELEASE_NAME="${RELEASE_NAME:-medgemma-official}"
+if [ -z "${API_KEY:-}" ]; then
+  echo "API_KEY must be set. Execute:" >&2
+  echo "read -rs API_KEY && export API_KEY" >&2
+  echo "And then type a desired key." >&2
+  exit 1
+fi
+IMAGE_REPOSITORY="${IMAGE_REPOSITORY:-vllm/vllm-openai}"
+IMAGE_TAG="${IMAGE_TAG:-v0.18.0}"
+GPU_MEMORY_UTILIZATION="${GPU_MEMORY_UTILIZATION:-0.40}"
+MAX_MODEL_LEN="${MAX_MODEL_LEN:-8192}"
+CPU_REQUEST="${CPU_REQUEST:-8}"
+MEMORY_REQUEST="${MEMORY_REQUEST:-48Gi}"
+GPU_COUNT="${GPU_COUNT:-1}"
+PVC_STORAGE="${PVC_STORAGE:-80Gi}"
+INGRESS_CLASS="${INGRESS_CLASS:-nginx}"
+
+if [ -z "${HF_TOKEN:-}" ] && [ -f "${SCRIPT_DIR}/.hf_token" ]; then
+  HF_TOKEN="$(cat "${SCRIPT_DIR}/.hf_token")"
+fi
+
+if [ -z "${HF_TOKEN:-}" ]; then
+  echo "HF_TOKEN is required for ${MODEL_NAME}." >&2
+  echo "Set HF_TOKEN in the environment or create ${SCRIPT_DIR}/.hf_token." >&2
+  exit 1
+fi
+
+need() { command -v "$1" >/dev/null 2>&1 || { echo "Missing dependency: $1" >&2; exit 1; }; }
+need kubectl
+need helm
+
+NAMESPACE="${NAMESPACE:-$(kubectl config view --minify -o jsonpath='{..namespace}' 2>/dev/null || true)}"
+if [ -z "${NAMESPACE}" ]; then
+  NAMESPACE="llm"
+fi
+
+SECRET_NAME="${RELEASE_NAME}-auth"
+SERVICE_NAME="${RELEASE_NAME}-${MODEL_ENTRY_NAME}-engine-service"
+DEPLOY_LABEL_MODEL="${MODEL_ENTRY_NAME}"
+INGRESS_NAME="${RELEASE_NAME}-api-ingress"
+
+echo "==> Runtime: vLLM (official helm chart)"
+echo "==> Namespace: ${NAMESPACE}"
+echo "==> Release: ${RELEASE_NAME}"
+echo "==> API host: ${API_HOST}"
+echo "==> Model: ${MODEL_NAME}"
+echo "==> Model entry name: ${MODEL_ENTRY_NAME}"
+echo "==> Image: ${IMAGE_REPOSITORY}:${IMAGE_TAG}"
+echo "==> Max model length: ${MAX_MODEL_LEN}"
+echo "==> GPU memory utilization: ${GPU_MEMORY_UTILIZATION}"
+echo
+
+kubectl get ns "${NAMESPACE}" >/dev/null 2>&1 || kubectl create ns "${NAMESPACE}"
+
+helm repo add vllm https://vllm-project.github.io/production-stack >/dev/null 2>&1 || true
+helm repo update >/dev/null 2>&1
+
+cat <<EOF | kubectl -n "${NAMESPACE}" apply -f -
+apiVersion: v1
+kind: Secret
+metadata:
+  name: ${SECRET_NAME}
+type: Opaque
+stringData:
+  HF_TOKEN: ${HF_TOKEN}
+  VLLM_API_KEY: ${API_KEY}
+EOF
+
+VALUES_FILE="$(mktemp)"
+cleanup() {
+  rm -f "${VALUES_FILE}"
+}
+trap cleanup EXIT
+
+cat > "${VALUES_FILE}" <<EOF
+servingEngineSpec:
+  runtimeClassName: nvidia
+  vllmApiKey:
+    secretName: ${SECRET_NAME}
+    secretKey: VLLM_API_KEY
+  modelSpec:
+    - name: ${MODEL_ENTRY_NAME}
+      repository: ${IMAGE_REPOSITORY}
+      tag: ${IMAGE_TAG}
+      modelURL: ${MODEL_NAME}
+      replicaCount: 1
+      requestCPU: ${CPU_REQUEST}
+      requestMemory: "${MEMORY_REQUEST}"
+      requestGPU: ${GPU_COUNT}
+      pvcStorage: "${PVC_STORAGE}"
+      vllmConfig:
+        dtype: bfloat16
+        maxModelLen: ${MAX_MODEL_LEN}
+        gpuMemoryUtilization: ${GPU_MEMORY_UTILIZATION}
+        enablePrefixCaching: true
+        enableChunkedPrefill: true
+        maxNumSeqs: 4
+        extraArgs:
+          - --mm-processor-cache-gb
+          - "1"
+          - --max-num-batched-tokens
+          - "2048"
+      hf_token:
+        secretName: ${SECRET_NAME}
+        secretKey: HF_TOKEN
+routerSpec:
+  enableRouter: false
+EOF
+
+KUBECONFIG="${KUBECONFIG:-}" helm upgrade --install "${RELEASE_NAME}" vllm/vllm-stack \
+  --namespace "${NAMESPACE}" \
+  -f "${VALUES_FILE}" \
+  --skip-crds \
+  --wait --timeout=20m
+
+cat <<EOF | kubectl -n "${NAMESPACE}" apply -f -
+apiVersion: networking.k8s.io/v1
+kind: Ingress
+metadata:
+  name: ${INGRESS_NAME}
+  annotations:
+    kubernetes.io/ingress.class: ${INGRESS_CLASS}
+    nginx.ingress.kubernetes.io/proxy-body-size: "0"
+    nginx.ingress.kubernetes.io/proxy-read-timeout: "600"
+    nginx.ingress.kubernetes.io/proxy-send-timeout: "600"
+    nginx.ingress.kubernetes.io/proxy-connect-timeout: "600"
+    nginx.ingress.kubernetes.io/proxy-buffering: "off"
+spec:
+  rules:
+  - host: ${API_HOST}
+    http:
+      paths:
+      - path: /
+        pathType: Prefix
+        backend:
+          service:
+            name: ${SERVICE_NAME}
+            port:
+              number: 80
+EOF
+
+echo
+echo "==> Pods:"
+kubectl -n "${NAMESPACE}" get pods -o wide
+echo
+echo "==> Services:"
+kubectl -n "${NAMESPACE}" get svc -o wide
+echo
+echo "==> Ingress:"
+kubectl -n "${NAMESPACE}" get ingress -o wide
+echo
+echo "==> Waiting for MedGemma pod readiness..."
+kubectl -n "${NAMESPACE}" wait --for=condition=ready pod \
+  -l "model=${DEPLOY_LABEL_MODEL},helm-release-name=${RELEASE_NAME}" \
+  --timeout=900s
+echo
+echo "==> Ready"
+echo "Base URL: https://${API_HOST}/v1"
+echo "Model: ${MODEL_NAME}"
+echo "Example:"
+echo "  curl https://${API_HOST}/v1/models -H 'Authorization: Bearer ${API_KEY}'"