diff --git a/.github/workflows/cogstack-cohorter-docker.yml b/.github/workflows/cogstack-cohorter-docker.yml
new file mode 100644
index 0000000..d648ef6
--- /dev/null
+++ b/.github/workflows/cogstack-cohorter-docker.yml
@@ -0,0 +1,71 @@
+name: ci-build-cohorter
+
+on:
+  push:
+    branches: [main]
+    tags:
+      - "cohorter-v*.*.*" # e.g., cohorter-v1.2.3
+    paths:
+      - "cogstack-cohorter/**"
+      - ".github/workflows/cogstack-cohorter-docker**"
+  pull_request:
+    paths:
+      - "cogstack-cohorter/**"
+      - ".github/workflows/cogstack-cohorter-docker**"
+
+jobs:
+  docker:
+    runs-on: ubuntu-latest
+    if: github.event.pull_request.user.login != 'dependabot[bot]' && github.repository == 'CogStack/cogstack-platform'
+    strategy:
+      matrix:
+        include:
+          - name: nl2dsl
+            context: cogstack-cohorter/NL2DSL
+            dockerfile: cogstack-cohorter/NL2DSL/Dockerfile
+            image: cogstacksystems/cogstack-cohorter-nl2dsl
+
+          - name: webapp
+            context: cogstack-cohorter/WebAPP
+            dockerfile: cogstack-cohorter/WebAPP/Dockerfile
+            image: cogstacksystems/cogstack-cohorter-webapp
+
+    steps:
+      - name: Checkout branch
+        uses: actions/checkout@v4
+
+      - name: Login to DockerHub
+        uses: docker/login-action@v3
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_TOKEN }}
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Extract metadata (tags, labels) for Docker
+        id: meta
+        uses: docker/metadata-action@v5
+        with:
+          images: ${{ matrix.image }}
+          tags: |
+            # latest tag on the default branch
+            type=raw,value=latest,enable={{is_default_branch}}
+            # Include all default tags
+            type=schedule
+            type=ref,event=branch
+            type=ref,event=tag
+            type=ref,event=pr
+            type=sha
+
+      - name: Build and push Docker image
+        id: push
+        uses: docker/build-push-action@v6
+        with:
+          context: ${{ matrix.context }}
+          file: ${{ matrix.dockerfile }}
+          push: ${{ github.event_name != 'pull_request' }}
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}
+          cache-from: type=registry,ref=${{ matrix.image }}:buildcache
+          cache-to: type=registry,ref=${{ matrix.image }}:buildcache,mode=max
diff --git a/cogstack-cohorter/WebAPP/.dockerignore b/cogstack-cohorter/WebAPP/.dockerignore
new file mode 100644
index 0000000..087b685
--- /dev/null
+++ b/cogstack-cohorter/WebAPP/.dockerignore
@@ -0,0 +1,7 @@
+node_modules/
+client/node_modules/
+client-react/node_modules/
+server/node_modules/
+
+# Data files are never baked into the image — supply them via volume mount at runtime
+server/data/
diff --git a/cogstack-cohorter/WebAPP/Dockerfile b/cogstack-cohorter/WebAPP/Dockerfile
index 421a824..7a10318 100644
--- a/cogstack-cohorter/WebAPP/Dockerfile
+++ b/cogstack-cohorter/WebAPP/Dockerfile
@@ -1,13 +1,13 @@
 FROM node:latest
 WORKDIR /usr/src/app
-COPY . .
 
-RUN cd /usr/src/app/server/data && tar xzvf snomed_terms_data.tar.gz
+COPY . .
+RUN cd /usr/src/app/client-react && npm install && npm run build
+RUN cd /usr/src/app/server && npm install
 
-ARG random
-RUN if [ "$random" = "true" ] ; then cd /usr/src/app/server/data && node --max-old-space-size=32768 gen_random_data.js ; fi
+COPY entrypoint.sh /entrypoint.sh
+RUN sed -i 's/\r$//' /entrypoint.sh && chmod +x /entrypoint.sh
 
-RUN cd /usr/src/app/server && npm install
 EXPOSE 3000
 WORKDIR /usr/src/app/server
-CMD ["node", "--max-old-space-size=32768", "server.js"]
+ENTRYPOINT ["/entrypoint.sh"]
diff --git a/cogstack-cohorter/WebAPP/entrypoint.sh b/cogstack-cohorter/WebAPP/entrypoint.sh
new file mode 100644
index 0000000..b5db8b6
--- /dev/null
+++ b/cogstack-cohorter/WebAPP/entrypoint.sh
@@ -0,0 +1,28 @@
+#!/bin/sh
+set -e
+
+DATA_DIR=/usr/src/app/server/data
+
+# ── Step 1: extract archive if JSON data isn't already present ────────────────
+if [ ! -f "$DATA_DIR/snomed_terms.json" ]; then
+    if [ -f "$DATA_DIR/snomed_terms_data.tar.gz" ]; then
+        echo "[webapp] Extracting SNOMED data archive..."
+        tar xzvf "$DATA_DIR/snomed_terms_data.tar.gz" -C "$DATA_DIR"
+    else
+        echo "[webapp] ERROR: No data found at $DATA_DIR." >&2
+        echo "[webapp] Mount a directory containing snomed_terms.json (and related files)" >&2
+        echo "[webapp] or snomed_terms_data.tar.gz via a Docker volume:" >&2
+        echo "[webapp]   -v /your/data:/usr/src/app/server/data" >&2
+        exit 1
+    fi
+fi
+
+# ── Step 2 (optional): generate random patient data ───────────────────────────
+# Set RANDOM_DATA=true in the container environment to generate synthetic data.
+if [ "${RANDOM_DATA}" = "true" ]; then
+    echo "[webapp] Generating random demo patient data..."
+    node --max-old-space-size=32768 /usr/src/app/server/gen_random_data.js
+fi
+
+# ── Step 3: start the server ──────────────────────────────────────────────────
+exec node --max-old-space-size=32768 server.js
diff --git a/cogstack-cohorter/WebAPP/server/gen_random_data.js b/cogstack-cohorter/WebAPP/server/gen_random_data.js
new file mode 100644
index 0000000..d242e75
--- /dev/null
+++ b/cogstack-cohorter/WebAPP/server/gen_random_data.js
@@ -0,0 +1,66 @@
+// Use this script to generate random data for the app
+// Run this script with the command:
+// node --max-old-space-size=32768 gen_random_data.js
+console.log('Generating random data')
+const fs = require('fs');
+const snomed_terms = require('./data/snomed_terms.json');
+
+// Returns a random integer between min (inclusive) and max (inclusive).
+function random_int(min, max) {
+    min = Math.ceil(min);
+    max = Math.floor(max);
+    return Math.floor(Math.random() * (max - min + 1)) + min;
+}
+
+const sex_id2code = ['Male', 'Female', 'Unknown']
+const eth_id2code = ['Asian', 'Black', 'White', 'Mixed', 'Other', 'Unknown'];
+let ptt2age = {};
+let ptt2sex = {};
+let ptt2eth = {};
+let ptt2dod = {};
+let cui2ptt_pos = {};
+let cui2ptt_tsp = {};
+
+let ptt_num = 100000;
+let max_ptt = 1000; // max. number of ptt a term can have
+let max_age = 100;
+let die_pct = 10; // percentage of died ptt = 1 / die_pct
+
+// generate ptt_num random patient data
+for (let i=0;i<ptt_num;i++) {
+    ptt2age[i] = random_int(0,max_age);
+    ptt2sex[i] = sex_id2code[random_int(0,sex_id2code.length-1)];
+    ptt2eth[i] = eth_id2code[random_int(0,eth_id2code.length-1)];
+    ptt2dod[i] = random_int(0,die_pct) == 0 ? random_int(Math.floor(Date.now()/1000) - (60*60*24*365*10), Math.floor(Date.now()/1000)) : 0;
+    if (i%100000 == 0) console.log('ptt:', i, `${Math.floor((i/ptt_num)*100)}%`);
+}
+
+// for each snomed terms, generate some random mention data
+for (let i=0;i<snomed_terms.length;i++) {
+    if (snomed_terms[i]['str'].search('(disorder)')==-1 && snomed_terms[i]['str'].search('(finding)')==-1 &&
+        snomed_terms[i]['str'].search('(procedure)')==-1 && snomed_terms[i]['str'].search('(substance)')==-1)
+        continue;
+    let picked = {};
+    cui2ptt_pos[i] = {};
+    cui2ptt_tsp[i] = {};
+    for (let j=0;j<random_int(0,max_ptt);j++) {
+        let ptt = random_int(0, ptt_num-1);
+        while (picked[ptt]) ptt = random_int(0, ptt_num-1);
+        picked[ptt] = true;
+        cui2ptt_pos[i][ptt] = random_int(1,100);
+        cui2ptt_tsp[i][ptt] = random_int(Math.floor(Date.now()/1000) - (60*60*24*365*10), Math.floor(Date.now()/1000));
+    }
+    if (i%100000 == 0) console.log('men:', i, `${Math.floor((i/snomed_terms.length)*100)}%`);
+}
+
+// write to files
+console.log('Writing to files...')
+fs.writeFileSync('data/ptt2age.json', JSON.stringify(ptt2age));
+fs.writeFileSync('data/ptt2sex.json', JSON.stringify(ptt2sex));
+fs.writeFileSync('data/ptt2eth.json', JSON.stringify(ptt2eth));
+fs.writeFileSync('data/ptt2dod.json', JSON.stringify(ptt2dod));
+const pos_out = fs.createWriteStream('data/cui2ptt_pos.jsonl', {flags: 'w'});
+const tsp_out = fs.createWriteStream('data/cui2ptt_tsp.jsonl', {flags: 'w'});
+Object.keys(cui2ptt_pos).forEach( k => { pos_out.write(`{"${snomed_terms[k]['cui']}":` + JSON.stringify(cui2ptt_pos[k]) + '}\n'); });
+Object.keys(cui2ptt_tsp).forEach( k => { tsp_out.write(`{"${snomed_terms[k]['cui']}":` + JSON.stringify(cui2ptt_tsp[k]) + '}\n'); });
+console.log('Finished generating random data')
\ No newline at end of file
diff --git a/cogstack-cohorter/WebAPP/server/server.js b/cogstack-cohorter/WebAPP/server/server.js
index e1157b8..744c597 100644
--- a/cogstack-cohorter/WebAPP/server/server.js
+++ b/cogstack-cohorter/WebAPP/server/server.js
@@ -28,7 +28,7 @@ app.use(session({
 }))
 
 
-const NL2DSL_SERVER = process.env.NL2DSL_SERVER || "http://localhost:4000/api/compile";
+const NL2DSL_SERVER = process.env.NL2DSL_SERVER || "http://localhost:3002/api/compile";
 
 let port = process.env.PORT || 3000;
 console.log('Loading data...');
@@ -181,7 +181,7 @@ for (let i=0;i<all_ptt_cnt;i++) ptt2cui_tsp_arr[i] = {};
     }
     console.log('Finished reading cui2ptt_tsp');
     console.timeEnd('cui2ptt_tsp');
-    console.log('Loading data...Finsihed');
+    console.log('Finished loading data');
     console.log(`Access the app on port ${port}`);
 })();
 //========================================================
diff --git a/cogstack-cohorter/docker-compose.yml b/cogstack-cohorter/docker-compose.yml
index 83ac9bf..65ca0fe 100644
--- a/cogstack-cohorter/docker-compose.yml
+++ b/cogstack-cohorter/docker-compose.yml
@@ -8,6 +8,22 @@ services:
       - ollama:/root/.ollama
     restart: unless-stopped
 
+  ollama-init:
+    image: ollama/ollama:latest
+    container_name: ollama-pull
+    entrypoint: >
+      sh -c "
+        until ollama list > /dev/null 2>&1; do
+          echo '[ollama-pull] Waiting for Ollama...'; sleep 2;
+        done &&
+        ollama pull ${OLLAMA_MODEL:-gpt-oss:20b}
+      "
+    environment:
+      OLLAMA_HOST: "http://ollama:11434"
+    depends_on:
+      - ollama
+    restart: "no"
+
   medcat:
     image: cogstacksystems/medcat-service:latest
     container_name: cohorter-medcat
@@ -32,20 +48,25 @@ services:
       MEDCAT_URL: "http://cohorter-medcat:5000"
       ALLOW_ORIGINS: "*"
     depends_on:
-      - ollama
+      - ollama-init
       - medcat
     restart: unless-stopped
 
   webapp:
     build:
       context: ./WebAPP
-      args:
-        random: "false"   # set to "true" to run gen_random_data.js during build
     container_name: cohorter-webapp
     ports:
       - "3000:3000"
     environment:
-      NL2DSL_URL: "http://cohorter-nl2dsl:3002"
+      NL2DSL_SERVER: "http://cohorter-nl2dsl:3002/api/compile"
+      # Set to "true" to generate synthetic patient data on startup.
+      RANDOM_DATA: "true"
+    volumes:
+      # Mount your data directory here. It should contain either:
+      #   - snomed_terms_data.tar.gz (will be auto-extracted on startup), or
+      #   - pre-extracted files: snomed_terms.json (SNOMED term lookup by Concept Unique Identifier) and cui_pt2ch.json (SNOMED CT ontology hierachy map).
+      - ./WebAPP/server/data:/usr/src/app/server/data
     depends_on:
       - medcat
       - nl2dsl