diff --git a/.env.example b/.env.example
index a79422d..cfb4d1d 100644
--- a/.env.example
+++ b/.env.example
@@ -2,6 +2,24 @@ BOOKSTACK_TOKEN_ID=votre_token_bookstack
 BOOKSTACK_TOKEN_SECRET=votre_secret_token
 BOOKSTACK_URL=http://localhost:6875
 
-MODEL_LLM=mon-model-IA
-MODEL_EMBEDDING=mon-model-embedding
-OPENAI_API_KEY=myapikey
\ No newline at end of file
+OPENAI_API_KEY=sk-proj-your-api-key
+MODEL_LLM=myllmmodel
+MODEL_EMBEDDING=myembeddingmodel
+
+COLLECTION_NAME=collection_test
+VECTOR_STORE_DIR=vector-store
+TOP_K=4
+BACKEND_MODEL_ID=chatbot-rag
+SYSTEM_PROMPT=You are a helpful assistant for internal documentation.
+
+ANYLLM_ADMIN_PASSWORD=your_anythingllm_password
+JWT_SECRET=your-random-secret-key
+
+# Network binding / ports
+CHATBOT_API_BIND_IP=1.1.1.1
+CHATBOT_API_HOST_PORT=CHATPORT
+CHATBOT_API_INTERNAL_PORT=CHATPORT
+
+ANYTHINGLLM_BIND_IP=1.1.1.1
+ANYTHINGLLM_HOST_PORT=PORT
+ANYTHINGLLM_CONTAINER_PORT=PORT
diff --git a/.github/workflows/ghcr.yaml b/.github/workflows/ghcr.yaml
new file mode 100644
index 0000000..6abb288
--- /dev/null
+++ b/.github/workflows/ghcr.yaml
@@ -0,0 +1,31 @@
+name: Create and publish all Docker images
+
+on:
+  push:
+    branches: ["main","chore/github-registry"]
+
+jobs:
+  build-docker-image:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Login to GHCR
+        run: echo "${{ secrets.GITHUB_TOKEN }}" | docker login ghcr.io -u "${{ github.actor }}" --password-stdin
+
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v3
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Build and publish Chabot Docker Image
+        uses: docker/build-push-action@v6
+        with:
+          context: .
+          file: Dockerfile.api
+          push: true
+          tags: |
+            ghcr.io/xpeho/chatbot-api:latest
+            ghcr.io/xpeho/chatbot-api:${{ github.sha }}
diff --git a/Dockerfile.api b/Dockerfile.api
new file mode 100644
index 0000000..9b42bc2
--- /dev/null
+++ b/Dockerfile.api
@@ -0,0 +1,16 @@
+FROM python:3.11-slim
+
+WORKDIR /app
+
+ENV PYTHONDONTWRITEBYTECODE=1
+ENV PYTHONUNBUFFERED=1
+
+COPY requirements.txt /app/requirements.txt
+RUN pip install --no-cache-dir -r /app/requirements.txt
+
+COPY app /app/app 
+COPY vector-store /app/vector-store
+
+EXPOSE 8000
+
+CMD ["sh", "-lc", "uvicorn app.main:app --host 0.0.0.0 --port ${CHATBOT_API_INTERNAL_PORT}"]
diff --git a/README.md b/README.md
index 85712bd..e9c5b81 100644
--- a/README.md
+++ b/README.md
@@ -1,45 +1,108 @@
 # Chatbot
 
-## Export BookStack pages to PDF
+This project provides:
+- BookStack page export to PDF
+- Vector store indexing (Chroma)
+- A RAG chatbot API (OpenAI-compatible endpoints)
+- An Open WebUI interface connected to the RAG API
 
-### Prerequisites
-- Python 3.9+
-- BookStack API token (token id + token secret)
-- BookStack URL
+## Prerequisites
+
+- Python 3.11+
+- Docker + Docker Compose
+- OpenAI API key with active billing/quota
+- BookStack API token (for export only)
+
+## Environment setup
+
+Create `.env` from `.env.example` and fill values:
+
+```env
+OPENAI_API_KEY=sk-proj-your-api-key
+MODEL_LLM=yout_model
+MODEL_EMBEDDING=your_embedding_model
+
+BOOKSTACK_URL=your_url
+BOOKSTACK_TOKEN_ID=your_bookstack_token_id
+BOOKSTACK_TOKEN_SECRET=your_bookstack_token_secret
+
+CHATBOT_API_BIND_IP=IP
+CHATBOT_API_HOST_PORT=PORT
+CHATBOT_API_INTERNAL_PORT=PORT
+ANYTHINGLLM_BIND_IP=IP
+ANYTHINGLLM_HOST_PORT=PORT
+```
+
+## Python setup
 
-### Setup
-1. Create a virtual environment
 ```bash
 python -m venv .venv
 source .venv/bin/activate
+pip install -r requirements.txt
 ```
 
-2. Install dependencies
+## 1) Export BookStack pages as PDFs
+
 ```bash
-pip install -r requirements.txt
+python export_pages.py
 ```
 
-3. Create a `.env` file (or update it) at the project root:
-```env
-BOOKSTACK_URL=your_bookstack_url
-BOOKSTACK_TOKEN_ID=your_token_id
-BOOKSTACK_TOKEN_SECRET=your_token_secret
-```
+PDF files are written to `exports/`.
+
+## 2) Build / refresh vector store
 
-## To run a notebook
 ```bash
-pip install jupyter ipykernel
-python -m ipykernel install --user --name chatbot-venv --display-name "Python (chatbot-venv)"
-jupyter notebook
+python reload_vector_store.py
 ```
 
-### Run the export
+This indexes documents from `exports/` into `vector-store/`.
+
+## 3) Run chatbot API + Open WebUI
+
 ```bash
-python export_pages.py
+docker compose up --build
 ```
 
-PDFs will be saved to `exports/`
+Access:
+- AnythingLLM: `http://localhost:${ANYTHINGLLM_HOST_PORT}` (default: `3001`)
+- Chatbot API health: `http://localhost:${CHATBOT_API_HOST_PORT}/health` (default: `8000`)
+
+Open WebUI is configured to call the local chatbot backend through OpenAI-compatible routes:
+- `GET /v1/models`
+- `POST /v1/chat/completions`
+
+## 4) Expose AnythingLLM on internet (secure minimal setup)
+
+Files:
+- `docker-compose.anythingllm.secure.yml`
+- `deploy/Caddyfile`
+
+This stack puts Caddy in front of AnythingLLM:
+- HTTPS with automatic TLS certificates
+- HTTP Basic Auth at proxy level
+- AnythingLLM not exposed directly (internal only)
+
+Steps:
+
+1. Create env file:
+
+docker compose up
 
-# Pricing OPenAI
-Pour un rechargement de base on utilise en moyenne 67 000 tokens soit 1/7 centimes.
-Une question côut 1/40 centimes (une toute simple).
+Security notes:
+- Keep AnythingLLM built-in auth enabled (multi-user recommended for internet exposure).
+- Disable public signup inside AnythingLLM unless explicitly needed.
+- Keep `OPENAI_API_KEY` only inside server env, never in frontend code.
+
+## 5) Automatize the vector store reload : 
+Create a job to automatically update the vectore store weekly : 
+CRON_TZ=Europe/Paris
+0 7 * * 1 chatbot/reload_job.sh >> chatbot/reindex.log 2>&1
+
+
+## Optional: notebooks
+
+```bash
+pip install jupyter ipykernel
+python -m ipykernel install --user --name chatbot-venv --display-name "Python (chatbot-venv)"
+jupyter notebook
+```
diff --git a/app/__init__.py b/app/__init__.py
new file mode 100644
index 0000000..8b13789
--- /dev/null
+++ b/app/__init__.py
@@ -0,0 +1 @@
+
diff --git a/app/main.py b/app/main.py
new file mode 100644
index 0000000..32a1581
--- /dev/null
+++ b/app/main.py
@@ -0,0 +1,100 @@
+#!/usr/bin/env python3
+import json
+import os
+import time
+import uuid
+from typing import Any
+
+from fastapi import FastAPI, HTTPException
+from fastapi.responses import StreamingResponse
+from pydantic import BaseModel
+
+BACKEND_MODEL_ID = os.getenv("BACKEND_MODEL_ID", "chatbot-rag")
+
+
+class ChatMessage(BaseModel):
+    role: str
+    content: Any
+
+
+class ChatCompletionRequest(BaseModel):
+    model: str | None = None
+    messages: list[ChatMessage]
+    temperature: float | None = None
+    stream: bool | None = False
+    max_tokens: int | None = None
+
+
+app = FastAPI(title="Chatbot API", version="0.1.0")
+
+
+@app.get("/v1/models")
+def list_models() -> dict[str, Any]:
+    return {
+        "object": "list",
+        "data": [
+            {
+                "id": BACKEND_MODEL_ID,
+                "object": "model",
+                "owned_by": "chatbot",
+            }
+        ],
+    }
+
+
+@app.post("/v1/chat/completions")
+def chat_completions(request: ChatCompletionRequest):
+    if not request.messages:
+        raise HTTPException(status_code=400, detail="messages is required")
+
+    answer = f"Reponse bidon API."
+    created = int(time.time())
+    completion_id = f"chatcmpl-{uuid.uuid4().hex[:24]}"
+    model_id = request.model or BACKEND_MODEL_ID
+
+    if request.stream:
+        ##réponse compatible ANYTHINGLLM
+        def event_stream():
+            chunk_1 = {
+                "id": completion_id,
+                "object": "chat.completion.chunk",
+                "created": created,
+                "model": model_id,
+                "choices": [{"index": 0, "delta": {"role": "assistant"}, "finish_reason": None}],
+            }
+            yield f"data: {json.dumps(chunk_1)}\n\n"
+
+            chunk_2 = {
+                "id": completion_id,
+                "object": "chat.completion.chunk",
+                "created": created,
+                "model": model_id,
+                "choices": [{"index": 0, "delta": {"content": answer}, "finish_reason": None}],
+            }
+            yield f"data: {json.dumps(chunk_2)}\n\n"
+
+            chunk_3 = {
+                "id": completion_id,
+                "object": "chat.completion.chunk",
+                "created": created,
+                "model": model_id,
+                "choices": [{"index": 0, "delta": {}, "finish_reason": "stop"}],
+            }
+            yield f"data: {json.dumps(chunk_3)}\n\n"
+            yield "data: [DONE]\n\n"
+
+        return StreamingResponse(event_stream(), media_type="text/event-stream")
+
+    return {
+        "id": completion_id,
+        "object": "chat.completion",
+        "created": created,
+        "model": model_id,
+        "choices": [
+            {
+                "index": 0,
+                "message": {"role": "assistant", "content": answer},
+                "finish_reason": "stop",
+            }
+        ],
+    }
diff --git a/docker-compose.yml b/docker-compose.yml
new file mode 100644
index 0000000..3d1c8e0
--- /dev/null
+++ b/docker-compose.yml
@@ -0,0 +1,44 @@
+services:
+  chatbot-api:
+    build:
+      context: .
+      dockerfile: Dockerfile.api
+    container_name: chatbot-api-local
+    env_file:
+      - .env
+    environment:
+      - VECTOR_STORE_DIR=/app/vector-store
+      - COLLECTION_NAME=collection_test
+      - TOP_K=4
+      - BACKEND_MODEL_ID=chatbot-rag
+    volumes:
+      - ./vector-store:/app/vector-store
+    ports:
+      - "${CHATBOT_API_BIND_IP}:${CHATBOT_API_HOST_PORT}:${CHATBOT_API_INTERNAL_PORT}"
+    restart: unless-stopped
+
+  anythingllm:
+    image: mintplexlabs/anythingllm:latest
+    container_name: anythingllm-local
+    depends_on:
+      - chatbot-api
+    cap_add:
+      - SYS_ADMIN
+    env_file:
+      - .env
+    environment:
+      - STORAGE_DIR=/app/server/storage
+      - GENERIC_OPEN_AI_BASE_PATH=http://chatbot-api:${CHATBOT_API_INTERNAL_PORT}/v1
+      - OPEN_AI_BASE_PATH=http://chatbot-api:${CHATBOT_API_INTERNAL_PORT}/v1
+      - AUTH_TOKEN=${ANYLLM_ADMIN_PASSWORD}
+      - JWT_SECRET=${JWT_SECRET}
+    extra_hosts:
+      - "host.docker.internal:host-gateway"
+    volumes:
+      - anythingllm-local-storage:/app/server/storage
+    ports:
+      - "${ANYTHINGLLM_BIND_IP}:${ANYTHINGLLM_HOST_PORT}:${ANYTHINGLLM_HOST_PORT}"
+    restart: unless-stopped
+
+volumes:
+  anythingllm-local-storage:
diff --git a/docker/docker-compose-prod.yml b/docker/docker-compose-prod.yml
new file mode 100644
index 0000000..9fe483a
--- /dev/null
+++ b/docker/docker-compose-prod.yml
@@ -0,0 +1,40 @@
+services:
+  chatbot-api:
+    image: ghcr.io/xpeho/chatbot-api:latest
+    pull_policy: always
+    env_file:
+      - .env
+    environment:
+      - VECTOR_STORE_DIR=/app/vector-store
+      - COLLECTION_NAME=collection_test
+      - TOP_K=4
+      - BACKEND_MODEL_ID=chatbot-rag
+    volumes:
+      - ./vector-store:/app/vector-store
+    ports:
+      - "${CHATBOT_API_BIND_IP}:${CHATBOT_API_HOST_PORT}:${CHATBOT_API_INTERNAL_PORT}"
+    restart: unless-stopped
+
+  anythingllm:
+    image: mintplexlabs/anythingllm:latest
+    container_name: anythingllm-local
+    depends_on:
+      - chatbot-api
+    cap_add:
+      - SYS_ADMIN
+    env_file:
+      - .env
+    environment:
+      - STORAGE_DIR=/app/server/storage
+      - GENERIC_OPEN_AI_BASE_PATH=http://chatbot-api:${CHATBOT_API_INTERNAL_PORT}/v1
+      - OPEN_AI_BASE_PATH=http://chatbot-api:${CHATBOT_API_INTERNAL_PORT}/v1
+      - AUTH_TOKEN=${ANYLLM_ADMIN_PASSWORD}
+      - JWT_SECRET=${JWT_SECRET}
+    volumes:
+      - anythingllm-local-storage:/app/server/storage
+    ports:
+      - "${ANYTHINGLLM_BIND_IP}:${ANYTHINGLLM_HOST_PORT}:${ANYTHINGLLM_HOST_PORT}"
+    restart: unless-stopped
+
+volumes:
+  anythingllm-local-storage:
diff --git a/reload_job.sh b/reload_job.sh
new file mode 100755
index 0000000..552aac3
--- /dev/null
+++ b/reload_job.sh
@@ -0,0 +1,23 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+PROJECT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+COMPOSE_FILE="$PROJECT_DIR/docker-compose.yml"
+LOCK_DIR="/tmp/chatbot-reindex-lock"
+
+# Prevent overlapping runs.
+if ! mkdir "$LOCK_DIR" 2>/dev/null; then
+  echo "[reindex] another run is already in progress"
+  exit 0
+fi
+trap 'rmdir "$LOCK_DIR"' EXIT
+
+cd "$PROJECT_DIR"
+
+echo "[reindex] starting export + vector store reload"
+docker compose -f "$COMPOSE_FILE" run --rm --no-deps --build \
+  -v "$PROJECT_DIR:/work" \
+  -w /work \
+  chatbot-api \
+  sh -lc "python export_pages.py && python reload_vector_store.py"
+echo "[reindex] done"
diff --git a/requirements.txt b/requirements.txt
index 9c37cbc..3417529 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,4 +1,6 @@
 python-dotenv>=1.0.0
+fastapi>=0.115.0
+uvicorn[standard]>=0.30.0
 langchain-openai
 langchain_chroma
 langchain-text-splitters
@@ -6,4 +8,4 @@ langchain-community
 langgraph
 pathlib
 tqdm
-pypdf
\ No newline at end of file
+pypdf