Refine architecture: Add Redis server, OmniRoute multi-port listeners (20128/20129/20132), safe SQLite backup, and OMNIROUTE_BASE_PATH

ARCHITECTURE.md

@@ -6,17 +6,21 @@ This document details the software architecture, network routing, multi-service
 
 ## 1. System Overview
 
-**Opencode-Cli** is a multi-service web container running on Debian Bookworm. It bundles five core microservices into a unified web application exposed via a single public port (`4096`):
-
-1. **FastAPI Reverse Proxy & Gateway Router**: Ingress routing, path-rewriting, HTML asset fixing, CORS, and WebSocket streaming.
-2. **Open WebUI**: Primary ChatGPT-style web interface. Pre-configured with OmniRoute as its backend OpenAI API provider.
-3. **OmniRoute AI Gateway**: Universal AI model routing, combo load balancing, fallback rules, circuit breakers, and OpenAI (`/v1`) & Gemini (`/v1beta`) API compatibility.
-4. **Jellyfin Media Server**: Media streaming center.
-5. **Telegram 5G Stream Proxy**: High-speed MTProto streaming daemon that converts Telegram files into `.strm` media files for Jellyfin.
+**Opencode-Cli** is an enterprise-grade multi-service web container running on Debian Bookworm Slim. It orchestrates six integrated services into a unified web application exposed via a single public ingress port (`4096`):
+
+1. **FastAPI Reverse Proxy & Gateway Router**: Ingress path-routing, CORS, WebSocket proxying, and service isolation.
+2. **Open WebUI**: Primary ChatGPT-style frontend pre-configured to use OmniRoute's dedicated API server (`127.0.0.1:20129/v1`).
+3. **OmniRoute v3.8.50 Multi-Port AI Gateway**:
+   - **Dashboard**: Port `20128` (Web UI & settings at `/omniroute`)
+   - **Dedicated API Server**: Port `20129` (OpenAI `/v1` & Gemini `/v1beta` compatibility)
+   - **Live WebSocket Server**: Port `20132` (Real-time monitoring)
+4. **Redis Cache & Rate Limiter**: Port `6379` (In-memory cache and distributed rate limiting for OmniRoute).
+5. **Jellyfin Media Server**: Port `8096` (Media streaming server at `/jellyfin`).
+6. **Telegram 5G Stream Proxy**: Port `8080` (High-speed MTProto streaming daemon at `/tg-stream`).
 
 ---
 
-## 2. System Architecture Diagram
+## 2. Production Architecture Diagram
 
 ```mermaid
 flowchart TD
@@ -24,19 +28,22 @@ flowchart TD
 
     subgraph Container ["🐳 Docker Container (Debian Bookworm Slim)"]
         Ingress -->|Path: / | OWUI[Open WebUI - Port 8098]
-        Ingress -->|Path: /omniroute | Omni[OmniRoute AI Gateway - Port 20128]
-        Ingress -->|Path: /v1 , /v1beta | Omni
-        Ingress -->|Path: /jellyfin | JF[Jellyfin Media Server - Port 8096]
-        Ingress -->|Path: /tg-stream | TG[Telegram 5G Stream Proxy - Port 8080]
-
-        OWUI -->|OpenAI API Format /v1| Omni
+        Ingress -->|Path: /omniroute/* | OmniDash[OmniRoute Dashboard - Port 20128]
+        Ingress -->|Path: /v1 , /v1beta | OmniAPI[OmniRoute Dedicated API - Port 20129]
+        Ingress -->|Path: /omniroute/live-ws/* | OmniWS[OmniRoute Live WS - Port 20132]
+        Ingress -->|Path: /jellyfin/* | JF[Jellyfin Media Server - Port 8096]
+        Ingress -->|Path: /tg-stream/* | TG[Telegram 5G Streamer - Port 8080]
+
+        OWUI -->|Internal OpenAI API /v1| OmniAPI
+        OmniDash -->|Distributed Cache & Limits| Redis[(Redis Server - Port 6379)]
+        OmniAPI -->|Distributed Cache & Limits| Redis
         TG -->|Auto-Generate .strm Files| JF
     end
 
-    subgraph Persistence ["💾 Persistence Architecture"]
-        Ext4["Local Container Ext4 (/root/.omniroute)"] <-->|Auto-Sync Daemon (25s)| HFData["HF FUSE Persistent Mount (/data)"]
-        Omni -->|POSIX SQLite Locks| Ext4
-        OWUI -->|User DB & Configs| HFData
+    subgraph Persistence ["💾 Persistence & Storage Model"]
+        Ext4["Local Container Ext4 (/root/.omniroute)"] <-->|SQLite .backup API (Every 5 mins)| HFData["HF FUSE Persistent Mount (/data)"]
+        OmniDash -->|Fast POSIX SQLite Locks| Ext4
+        OWUI -->|User Accounts & DB| HFData
         JF -->|Media & Metadata| HFData
     end
 ```
@@ -49,37 +56,42 @@ flowchart TD
 - **Port**: `4096` (Exposed as container entrypoint)
 - **Role**:
   - Handles single-port ingress on Hugging Face Spaces.
-  - Dynamically routes requests based on URL path prefixes.
-  - Rewrites HTML `<head>` tags and Next.js asset paths (`/_next`).
-  - Proxies HTTP requests and bidirectional WebSockets (`/ws/socket.io`).
+  - Namespace path-routing with strict boundary enforcement.
+  - Proxies HTTP requests and bidirectional WebSockets (`/ws/socket.io` & `/omniroute/live-ws`).
 
 ### 3.2 Open WebUI (`open-webui`)
 - **Port**: `8098`
 - **Path**: `/`
 - **Environment**:
-  - `OPENAI_API_BASE_URL="http://127.0.0.1:20128/v1"`
+  - `OPENAI_API_BASE_URL="http://127.0.0.1:20129/v1"`
   - `ENABLE_OPENAI_API="true"`
   - `RAG_AUTO_UPDATE_INDEX="false"`
-- **Optimization**: SentenceTransformer embeddings (`all-MiniLM-L6-v2`) are pre-baked into the Docker image cache to allow instant boot in **~1s**.
+- **Optimization**: SentenceTransformer embeddings (`all-MiniLM-L6-v2`) pre-cached in Docker image for 1-second startup.
+
+### 3.3 OmniRoute AI Gateway v3.8.50
+- **Ports**:
+  - **Dashboard**: `20128` (Path `/omniroute`)
+  - **Dedicated API**: `20129` (Endpoints `/v1`, `/v1beta`)
+  - **Live WebSocket**: `20132` (Path `/omniroute/live-ws`)
+- **Environment**:
+  - `OMNIROUTE_BASE_PATH="/omniroute"`
+  - `REDIS_URL="redis://127.0.0.1:6379"`
+  - `DATA_DIR="/root/.omniroute"`
+- **Auto-Fix**: `fix_omniroute.py` auto-resolves numeric migration version collisions (e.g. version 143 collisions) at build time and container boot.
 
-### 3.3 OmniRoute AI Gateway (`/omniroute`)
-- **Port**: `20128`
-- **Path**: `/omniroute` *(Redirects to `/omniroute/dashboard`)*
-- **API Endpoints**: `/v1` (OpenAI format), `/v1beta` (Gemini format), and 21 specific backend management routes (`/api/providers`, `/api/combos`, etc.).
-- **Engine**: Node.js 22 LTS + Next.js.
-- **Auto-Fix**: `fix_omniroute.py` auto-resolves numeric migration version collisions (e.g., version 143 collisions) at build time and container boot.
+### 3.4 Redis Server (`redis-server`)
+- **Port**: `6379`
+- **Role**: In-memory rate limiting and distributed caching for OmniRoute.
 
-### 3.4 Jellyfin Media Server (`jellyfin`)
+### 3.5 Jellyfin Media Server (`jellyfin`)
 - **Port**: `8096`
 - **Path**: `/jellyfin`
-- **Data Location**: `/data/jellyfin/data`, `/data/jellyfin/media/Movies`
 - **Engine**: Jellyfin 10.11 + FFmpeg 5.1.
 
-### 3.5 Telegram Direct Stream Proxy (`tg_streamer.py`)
+### 3.6 Telegram Direct Stream Proxy (`tg_streamer.py`)
 - **Port**: `8080`
 - **Path**: `/tg-stream`
 - **Engine**: Python Pyrogram MTProto Client.
-- **Function**: Automatically syncs Telegram channel media into `.strm` files inside Jellyfin's movie directory.
 
 ---
 
@@ -88,13 +100,10 @@ flowchart TD
 | Public Request Path | Target Upstream Port | Target Service | Function |
 | :--- | :--- | :--- | :--- |
 | `/` | `8098` | Open WebUI | Main Web UI |
-| `/openwebui/*` | `8098` | Open WebUI | WebUI Assets |
-| `/omniroute` | `20128` | OmniRoute AI Gateway | Management Dashboard |
-| `/omniroute/_next/*` | `20128` | OmniRoute AI Gateway | Next.js Static Assets |
-| `/v1/*` | `20128` | OmniRoute AI Gateway | OpenAI API Endpoint |
-| `/v1beta/*` | `20128` | OmniRoute AI Gateway | Gemini API Endpoint |
-| `/api/providers`, `/api/combos`, ... | `20128` | OmniRoute AI Gateway | OmniRoute Management APIs |
-| `/api/config`, `/api/v1/*` | `8098` | Open WebUI | Open WebUI Core APIs |
+| `/omniroute/*` | `20128` | OmniRoute Dashboard | Management Panel |
+| `/v1/*` | `20129` | OmniRoute API Server | Dedicated OpenAI API |
+| `/v1beta/*` | `20129` | OmniRoute API Server | Dedicated Gemini API |
+| `/omniroute/live-ws/*` | `20132` | OmniRoute WebSocket | Live Monitoring WS |
 | `/jellyfin/*` | `8096` | Jellyfin Media Server | Media Streaming |
 | `/tg-stream/*` | `8080` | TG Stream Proxy | Telegram Direct Streamer |
 
@@ -102,30 +111,8 @@ flowchart TD
 
 ## 5. Storage & Persistence Model
 
-```
-                    ┌──────────────────────────────────────────────┐
-                    │     Hugging Face Network Mount (/data)       │
-                    │   (FUSE Volume - Persistent Across Boots)   │
-                    └──────────────────────┬───────────────────────┘
-                                           │
-                        ┌──────────────────┴──────────────────┐
-                        │                                     │
-             ┌──────────▼──────────┐               ┌──────────▼──────────┐
-             │ /data/open-webui    │               │ /data/omniroute     │
-             │ (webui.db)          │               │ (storage.sqlite)    │
-             └─────────────────────┘               └──────────▲──────────┘
-                                                              │
-                                                   (Background Sync 25s)
-                                                              │
-                                                   ┌──────────┴──────────┐
-                                                   │ /root/.omniroute    │
-                                                   │ (Active Local Ext4) │
-                                                   └─────────────────────┘
-```
-
-- **HF FUSE Mount (`/data`)**: Persistent volume provided by Hugging Face Spaces.
-- **Local Ext4 Disk (`/root/.omniroute`)**: Stores OmniRoute's active `better-sqlite3` database file to prevent SQLite `disk I/O error` on FUSE mounts.
-- **Background Sync Daemon**: Runs every 25 seconds in `entrypoint.sh` to copy `/root/.omniroute/storage.sqlite` and `server.env` to `/data/omniroute/`.
+- **Active Execution (`/root/.omniroute/storage.sqlite`)**: Runs on local Ext4 container disk to support `better-sqlite3` POSIX file locking.
+- **Safe SQLite Backup (`/data/omniroute/storage.sqlite`)**: Background daemon uses `sqlite3 .backup` API every 5 minutes and on shutdown to write consistent snapshots to the HF FUSE mount.
 
 ---
 
@@ -133,15 +120,14 @@ flowchart TD
 
 ```
 .
-├── Dockerfile              # Multi-stage Docker image build
-├── ARCHITECTURE.md         # Architecture specification (this file)
+├── Dockerfile              # Multi-stage Docker image with Redis & pre-cached models
+├── ARCHITECTURE.md         # Architecture specification (this document)
 ├── README.md               # User quickstart guide
-├── entrypoint.sh           # Concurrent service startup & storage sync daemon
+├── entrypoint.sh           # Concurrent service startup, Redis daemon & SQLite backup
 ├── fix_omniroute.py        # Migration version collision auto-repair script
-├── nginx.conf              # Nginx configuration (optional reference)
+├── nginx.conf              # Nginx configuration reference
 ├── proxy.py                # FastAPI gateway entrypoint script
 ├── tg_streamer.py          # Pyrogram Telegram 5G stream proxy
-├── index.html              # Custom splash landing page
 └── gateway/                # FastAPI Modular Router Package
     ├── __init__.py
     ├── main.py             # App router aggregator

Dockerfile

@@ -24,6 +24,7 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
     g++ \
     sqlite3 \
     libsqlite3-dev \
+    redis-server \
  && rm -rf /var/lib/apt/lists/*
 
 # 2. Add Official Jellyfin APT Repository & Install Jellyfin Server, Web UI, and FFmpeg
@@ -44,12 +45,14 @@ RUN python3 -c "from sentence_transformers import SentenceTransformer; SentenceT
 RUN curl -fsSL https://nodejs.org/dist/v22.22.2/node-v22.22.2-linux-x64.tar.gz \
     | tar -xz -C /usr/local --strip-components=1
 
-# 5. Clone and Prepare OmniRoute AI Gateway (fast install, defer build to runtime dev server for instant Docker builds)
+# 5. Clone and Build OmniRoute AI Gateway (with OMNIROUTE_BASE_PATH=/omniroute)
 WORKDIR /omniroute
 ENV NEXT_TELEMETRY_DISABLED=1
 ENV OMNIROUTE_USE_TURBOPACK=0
-ENV NODE_OPTIONS="--max-old-space-size=2048"
+ENV NODE_OPTIONS="--max-old-space-size=2560"
 ENV DISABLE_ESLINT_PLUGIN=true
+ENV OMNIROUTE_BASE_PATH="/omniroute"
+ENV NEXT_PUBLIC_BASE_PATH="/omniroute"
 
 COPY fix_omniroute.py /fix_omniroute.py
 
@@ -57,7 +60,7 @@ RUN git clone --depth 1 https://github.com/diegosouzapw/OmniRoute.git /omniroute
  && python3 /fix_omniroute.py /omniroute \
  && npm install --legacy-peer-deps --no-audit --no-fund --prefer-offline \
  && npm rebuild better-sqlite3 --build-from-source \
- && NEXT_TELEMETRY_DISABLED=1 OMNIROUTE_USE_TURBOPACK=0 NODE_OPTIONS="--max-old-space-size=2560" npm run build || true
+ && OMNIROUTE_BASE_PATH="/omniroute" NEXT_PUBLIC_BASE_PATH="/omniroute" NEXT_TELEMETRY_DISABLED=1 OMNIROUTE_USE_TURBOPACK=0 NODE_OPTIONS="--max-old-space-size=2560" npm run build || true
 
 RUN mkdir -p /root/.cache /data/cache /data/omniroute /data/open-webui
 RUN chmod -R 777 /root/.cache /data/cache /omniroute

entrypoint.sh

@@ -32,24 +32,34 @@ fi
 # Clean up stale lock files
 rm -f /root/.omniroute/*.sqlite-wal /root/.omniroute/*.sqlite-shm /root/.omniroute/*.lock 2>/dev/null || true
 
-# Background sync process to preserve OmniRoute database and secrets to /data mount
+# Start Redis server for OmniRoute distributed caching and rate limiting
+echo "[INIT] Starting Redis server on port 6379..."
+redis-server --daemonize yes --bind 127.0.0.1 --port 6379 2>/dev/null || echo "[WARN] Redis server startup warning"
+
+# Safe SQLite Backup Process (Uses sqlite3 .backup API every 5 mins to prevent live DB corruption)
 (
     while true; do
-        sleep 25
-        if [ -f "/root/.omniroute/storage.sqlite" ]; then
-            cp -f /root/.omniroute/storage.sqlite /data/omniroute/storage.sqlite 2>/dev/null || true
-        fi
-        if [ -f "/root/.omniroute/server.env" ]; then
-            cp -f /root/.omniroute/server.env /data/omniroute/server.env 2>/dev/null || true
+        sleep 300
+        if [ -f "/root/.omniroute/storage.sqlite" ] && command -v sqlite3 >/dev/null 2>&1; then
+            sqlite3 /root/.omniroute/storage.sqlite ".backup /data/omniroute/storage.sqlite" 2>/dev/null || true
         fi
     done
 ) &
 
-# Start OmniRoute AI Gateway on Port 20128
-echo "[INIT] Starting OmniRoute AI Gateway on port 20128..."
+# Start OmniRoute AI Gateway (Multi-Port: Dashboard: 20128, Dedicated API: 20129, Live WS: 20132)
+echo "[INIT] Starting OmniRoute AI Gateway (Dashboard: 20128, API: 20129, WS: 20132)..."
 export PORT=20128
+export API_PORT=20129
+export LIVE_WS_PORT=20132
 export HOSTNAME="127.0.0.1"
 export DATA_DIR="/root/.omniroute"
+export REDIS_URL="redis://127.0.0.1:6379"
+export OMNIROUTE_BASE_PATH="/omniroute"
+export NEXT_PUBLIC_BASE_URL="${WEBUI_URL:-https://jishnupg-opencode-cli.hf.space}/omniroute"
+export OMNIROUTE_WS_BRIDGE_SECRET="${OMNIROUTE_WS_BRIDGE_SECRET:-opencode_ws_bridge_secret_2026}"
+export JWT_SECRET="${JWT_SECRET:-opencode_jwt_secret_2026}"
+export API_KEY_SECRET="${API_KEY_SECRET:-opencode_api_key_secret_2026}"
+
 if [ -d "/omniroute" ]; then
     cd /omniroute
     # Fix any upstream migration version collisions (e.g. version 143 collisions)
@@ -63,11 +73,11 @@ if [ -d "/omniroute" ]; then
         npm run start -- --port 20128 &
     else
         echo "[INIT] Building OmniRoute production assets to resolve WASM modules..."
-        NEXT_TELEMETRY_DISABLED=1 OMNIROUTE_USE_TURBOPACK=0 NODE_OPTIONS="--max-old-space-size=2560" npm run build 2>/dev/null || true
+        OMNIROUTE_BASE_PATH="/omniroute" NEXT_PUBLIC_BASE_PATH="/omniroute" NEXT_TELEMETRY_DISABLED=1 OMNIROUTE_USE_TURBOPACK=0 NODE_OPTIONS="--max-old-space-size=2560" npm run build 2>/dev/null || true
         if [ -d ".build/next" ] || [ -d ".next" ]; then
             npm run start -- --port 20128 &
         else
-            npx next dev --port 20128 -H 127.0.0.1 &
+            OMNIROUTE_BASE_PATH="/omniroute" npx next dev --port 20128 -H 127.0.0.1 &
         fi
     fi
     echo "[INIT] OmniRoute AI Gateway started in background."
@@ -102,11 +112,11 @@ mkdir -p /root/.cache /data/cache 2>/dev/null || true
 chmod -R 777 /root/.cache /data/cache 2>/dev/null || true
 
 # Pre-configure & Start Open WebUI on port 8098 (Mounted at / via FastAPI proxy)
-echo "[INIT] Starting Open WebUI on port 8098 pre-configured with OmniRoute..."
+echo "[INIT] Starting Open WebUI on port 8098 pre-configured with OmniRoute API..."
 if command -v open-webui >/dev/null 2>&1; then
-    # Tell Open WebUI its public-facing URL and point OpenAI API to OmniRoute port 20128
+    # Point OpenAI API to OmniRoute dedicated API port 20129
     export WEBUI_URL="${WEBUI_URL:-https://jishnupg-opencode-cli.hf.space}"
-    export OPENAI_API_BASE_URL="http://127.0.0.1:20128/v1"
+    export OPENAI_API_BASE_URL="http://127.0.0.1:20129/v1"
     export OPENAI_API_KEY="omniroute"
     export WEBUI_SECRET_KEY="${WEBUI_SECRET_KEY:-opencode_webui_jwt_secret_2026}"
     export ENABLE_OLLAMA_API="${ENABLE_OLLAMA_API:-false}"

gateway/omniroute.py

@@ -89,17 +89,24 @@ async def omniroute_main_route(request: Request, path: str = ""):
             
     return res
 
-# ── 2. OpenAI & Gemini Compatibility APIs ─────────────────────────────────────
+# ── 2. OpenAI & Gemini Compatibility APIs (Dedicated API Port 20129) ──────────
 @router.api_route("/v1", methods=["GET", "POST", "PUT", "DELETE", "PATCH", "HEAD", "OPTIONS"])
 @router.api_route("/v1/{path:path}", methods=["GET", "POST", "PUT", "DELETE", "PATCH", "HEAD", "OPTIONS"])
 async def omniroute_v1_api(request: Request, path: str = ""):
-    target = f"http://127.0.0.1:{OMNIROUTE_PORT}/v1/{path}" if path else f"http://127.0.0.1:{OMNIROUTE_PORT}/v1"
-    return await proxy_http_request(target, request, default_prefix="/omniroute")
+    target = f"http://127.0.0.1:{OMNIROUTE_API_PORT}/v1/{path}" if path else f"http://127.0.0.1:{OMNIROUTE_API_PORT}/v1"
+    return await proxy_http_request(target, request, default_prefix="")
 
 @router.api_route("/v1beta/{path:path}", methods=["GET", "POST", "PUT", "DELETE", "PATCH", "HEAD", "OPTIONS"])
 async def omniroute_v1beta_api(request: Request, path: str = ""):
-    target = f"http://127.0.0.1:{OMNIROUTE_PORT}/v1beta/{path}"
-    return await proxy_http_request(target, request, default_prefix="/omniroute")
+    target = f"http://127.0.0.1:{OMNIROUTE_API_PORT}/v1beta/{path}"
+    return await proxy_http_request(target, request, default_prefix="")
+
+# ── 3. OmniRoute Live Monitoring WebSocket (Dedicated WS Port 20132) ──────────
+@router.websocket("/omniroute/live-ws")
+@router.websocket("/omniroute/live-ws/{path:path}")
+async def omniroute_live_ws(websocket: Request, path: str = ""):
+    target_ws = f"ws://127.0.0.1:{OMNIROUTE_WS_PORT}/live-ws/{path}" if path else f"ws://127.0.0.1:{OMNIROUTE_WS_PORT}/live-ws"
+    await proxy_websocket_stream(websocket, target_ws)
 
 # ── 3. Next.js Static Asset Routing ──────────────────────────────────────────
 @router.api_route("/_next/{path:path}", methods=["GET", "POST", "PUT", "DELETE", "PATCH", "HEAD", "OPTIONS"])

gateway/utils.py

@@ -19,24 +19,28 @@ from fastapi.responses import StreamingResponse
 logger = logging.getLogger("GatewayUtils")
 
 # Port definitions
-WEBUI_PORT     = int(os.environ.get("WEBUI_PORT", 8098))
-OMNIROUTE_PORT = int(os.environ.get("OMNIROUTE_PORT", 20128))
-JELLYFIN_PORT  = int(os.environ.get("JELLYFIN_PORT", 8096))
-TG_PORT        = int(os.environ.get("TG_PORT", 8080))
-GATEWAY_PORT   = int(os.environ.get("PORT", 4096))
+WEBUI_PORT         = int(os.environ.get("WEBUI_PORT", 8098))
+OMNIROUTE_PORT     = int(os.environ.get("OMNIROUTE_PORT", 20128))
+OMNIROUTE_API_PORT = int(os.environ.get("OMNIROUTE_API_PORT", 20129))
+OMNIROUTE_WS_PORT  = int(os.environ.get("OMNIROUTE_WS_PORT", 20132))
+JELLYFIN_PORT      = int(os.environ.get("JELLYFIN_PORT", 8096))
+TG_PORT            = int(os.environ.get("TG_PORT", 8080))
+GATEWAY_PORT       = int(os.environ.get("PORT", 4096))
 
 PUBLIC_HOST   = os.environ.get("PUBLIC_HOST", "jishnupg-opencode-cli.hf.space")
 PUBLIC_ORIGIN = f"https://{PUBLIC_HOST}"
 
-_INTERNAL_PORTS = {GATEWAY_PORT, WEBUI_PORT, OMNIROUTE_PORT, JELLYFIN_PORT, TG_PORT}
+_INTERNAL_PORTS = {GATEWAY_PORT, WEBUI_PORT, OMNIROUTE_PORT, OMNIROUTE_API_PORT, OMNIROUTE_WS_PORT, JELLYFIN_PORT, TG_PORT}
 _PORT_STRIP_RE  = re.compile(r"(https?://[^/:]+):(" + "|".join(str(p) for p in _INTERNAL_PORTS) + r")")
 
 _PORT_PREFIX_MAP = {
-    GATEWAY_PORT:   "",
-    OMNIROUTE_PORT: "/omniroute",
-    WEBUI_PORT:     "/openwebui",
-    JELLYFIN_PORT:  "/jellyfin",
-    TG_PORT:        "/tg-stream",
+    GATEWAY_PORT:       "",
+    OMNIROUTE_PORT:     "/omniroute",
+    OMNIROUTE_API_PORT: "",
+    OMNIROUTE_WS_PORT:  "/live-ws",
+    WEBUI_PORT:         "/openwebui",
+    JELLYFIN_PORT:      "/jellyfin",
+    TG_PORT:            "/tg-stream",
 }
 
 # Shared Async HTTPX Client Instance