Interactive Deployment Studio

Download & Deploy Nomido-Photo-Backend

Get an instant, customized deployment script tuned to your host’s memory and CPU resources in three quick steps.

1

Select Deployment Media

Choose whether to deploy on a Docker server/NAS or on your personal desktop computer.

2

Select Your Configuration

Tunes CPU worker threads, RAM footprint, and InsightFace detection models (`low`, `medium`, `high`).

Mode: medium • AI Model: buffalo_m
3

Server (Docker) Setup & Download

Configuration: Medium (4 GB RAM) • Mode: MODE=medium • Model: buffalo_m • Threads: 2

Fast Setup — Ready-to-Copy Terminal Command
curl -sL https://photo.nomidocloud.com/install.sh | MODE=medium bash
Step 1: Run the Installer Command

Open your terminal and paste the curl command above. The installer will verify Docker and docker-compose availability.

Step 2: Wait for Container Initialization

Pre-built multi-arch images will be pulled from GitHub Container Registry (GHCR) and started automatically.

Step 3: Access the Web App

Open http://localhost:8080 (or your server's IP) to set up your admin account and start uploading photos.

Mobile Companion

Nomido for iOS & Android

Stream your photos, view face clusters, and auto-backup memories directly to your personal NAS with zero cloud intermediaries.

Nomido for iOS

Requires iOS 16.0+
TestFlight / App Store
  • Background background photo sync & Live Photos playback
  • End-to-end encrypted WebRTC remote connection
  • Local offline album caching & facial cluster browser
Scan to Install
App Store

Nomido for Android

Android 10.0+ (ARM64 / x86_64)
Google Play / APK
  • Fast camera roll auto-sync with Wi-Fi only option
  • Direct P2P streaming without open router ports
  • Full map view with offline reverse geocoded cities
Scan to Install
Google Play
Next Steps

Post-Installation Verification

Follow these 3 quick steps after your containers are healthy.

Recommended. No open ports, no domain: the Hub links your server automatically.

1. Sign In to the Hub

Open the Nomido-Photo app on your phone (App Store or Google Play) and sign in or create your free Hub account with email, Google or Apple.

2. Pick Your Server

The app scans your Wi-Fi and finds your server (or type its IP address). Select it and create your administrator account. Your Hub API key is generated and linked automatically: nothing to copy or paste.

3. You're Connected

Photos sync from anywhere over an encrypted peer-to-peer link (STUN / WebRTC). No open ports, no domain, no certificates. Your web gallery is available at photo.nomidocloud.com/<your-username>.

PRO / ADVANCED100% Local-Only Capable

Advanced Configuration & .env Reference

Line-by-Line Configuration Documentation

Step-by-step expandable parameter reference and complete environment variable manual for Docker Compose, homelab administrators, custom reverse proxies, and isolated air-gapped deployments.

No Cloud Account Required (Local-Only Architecture)

Nomido-Photo never phones home by default. If you prefer to expose your instance using your own domain, reverse proxy (Nginx, Traefik, Caddy), or a private VPN (WireGuard, Tailscale), simply omit HUB_API_KEY. The companion mobile app features a dedicated "Skip website configuration (Direct Register)" option to connect straight to your local IP or reverse proxy domain without creating a commercial account.

L3dbService RootPostgreSQL Vector Database Service
  db:

Primary relational and vector database storing user accounts, EXIF metadata, albums, and 512-dimensional CLIP/InsightFace embeddings.

L4imageContainer ImagePostgreSQL 15 + pgvector Image
    image: ankane/pgvector:v0.5.1-pg15

Pre-compiled PostgreSQL 15 image bundled with the pgvector extension for sub-millisecond cosine similarity search on faces and semantic queries.

L5restartLifecycleAutomatic Database Recovery
    restart: unless-stopped

Ensures the database container automatically restarts after host reboots or unexpected process termination unless explicitly stopped.

L6environmentEnvironmentDatabase Environment Block
    environment:

Configures initial superuser credentials and default database creation on first container boot.

L7POSTGRES_USEREnvironmentDatabase Role Username
      POSTGRES_USER: photo_user

Dedicated PostgreSQL role used by the Rust backend to execute parameterized SQLx queries.

L8POSTGRES_PASSWORDSecurity SecretDatabase Role Password
      POSTGRES_PASSWORD: secure_password_here

Secret password for photo_user. Replace with a strong 16+ byte random string in production.

L9POSTGRES_DBEnvironmentDatabase Schema Name
      POSTGRES_DB: photo_db

Name of the database automatically provisioned and migrated on startup.

L10volumesVolume MountDatabase Volume Mounts
    volumes:

Maps host storage into the container so database tables and vector indexes persist across container upgrades.

L11./data/postgresVolume MountPersistent PGData Directory
      - ./data/postgres:/var/lib/postgresql/data

Stores PostgreSQL data files in ./data/postgres on your host filesystem so state survives container destruction.

L12healthcheckHealthcheckReadiness Health Probe
    healthcheck:

Allows Docker to verify that PostgreSQL is accepting connections before starting dependent containers.

L13healthcheck.testHealthcheckpg_isready Probe Command
      test: ["CMD-SHELL", "pg_isready -U photo_user -d photo_db"]

Executes pg_isready inside the container to confirm the photo_db catalog is online and accepting connections.

L14healthcheck.intervalHealthcheckProbe Interval
      interval: 5s

Runs the database readiness check every 5 seconds.

L15healthcheck.timeoutHealthcheckProbe Timeout
      timeout: 5s

Marks a single health probe attempt as failed if it takes longer than 5 seconds.

L16healthcheck.retriesHealthcheckMaximum Consecutive Retries
      retries: 5

Allows up to 5 consecutive startup retries before marking the container unhealthy.

L17backendService RootRust Actix-Web API & Media Server
  backend:

Core photo engine handling authentication, media streaming, WebRTC P2P tunneling, EXIF parsing, and background AI job orchestration.

L18imageContainer ImageMulti-Arch Rust Backend Image
    image: ghcr.io/nomidocloud/nomido-photo-backend:latest

Official pre-built linux/amd64 and linux/arm64 image pulled from GitHub Container Registry.

L19restartLifecycleBackend Restart Policy
    restart: unless-stopped

Automatically restarts the API server on host reboot or crash.

L20depends_onLifecycleStartup Dependency Graph
    depends_on:

Prevents the backend from booting before its required database dependency is ready.

L21depends_on.dbLifecycleDatabase Dependency Target
      db:

Links startup order to the db service defined above.

L22conditionLifecycleWait for Healthy Database
        condition: service_healthy

Waits until pg_isready succeeds so SQLx database migrations run without connection race conditions.

L23environmentEnvironmentBackend Runtime Environment
    environment:

Injects database connection strings, cryptographic secrets, and hardware thread allocations.

L24DATABASE_URLSecurity SecretSQLx Connection URI
      DATABASE_URL: postgres://photo_user:secure_password_here@db:5432/photo_db

Connects the Rust backend to the db container over the internal Docker bridge network on port 5432 with pgvector extension enabled.

L25JWT_SECRETSecurity SecretCryptographic Session Signing Key
      JWT_SECRET: your_random_32_char_secret_key_here

256-bit secret key used to sign and verify user JWT access and refresh tokens. Generate with: openssl rand -hex 32.

L26AI_SERVICE_URLNetwork PortInternal AI Sidecar Endpoint
      AI_SERVICE_URL: http://ai-service:8000

Routes internal face detection, CLIP embedding, OCR, and geocoding requests to the ai-service container.

L27WORKER_THREADSHardware TunedHardware Thread Allocation (Medium Profile)
      WORKER_THREADS: 2

Tuned to 2 worker threads (4 GB RAM) for TOTAL_THREAD_NUMBER pool allocation matching your selected Medium hardware tier.

L28STORAGE_PATHEnvironmentInternal Media Vault Path
      STORAGE_PATH: /app/storage

Root directory path inside the container where original media, thumbnails, and cache are saved.

L29SERVER_PORTNetwork PortHTTP API Listener Port
      SERVER_PORT: 8080

HTTP port where the Rust Actix-Web backend listens for client API and WebSocket signaling connections.

L30APP_URLNetwork PortPublic Backend Base URL
      APP_URL: http://localhost:8080

Public URL of the backend instance, used for CORS headers and public sharing links.

L31HUB_API_KEYOptional RelayEncrypted P2P Relay Key
      HUB_API_KEY: optional_p2p_relay_key

Encrypted P2P relay connection key. Omit entirely for 100% offline or air-gapped local-only operation without any cloud dependencies.

L32LOG_LEVELEnvironmentLogging Verbosity Level
      LOG_LEVEL: info

Controls backend log verbosity: trace, debug, info, warn, or error.

L33volumesVolume MountMedia Storage Volume Mapping
    volumes:

Mounts your physical host drive or NAS RAID array into the backend container.

L34./data/storageVolume MountHost Photo Library Mount
      - ./data/storage:/app/storage

Persists all uploaded photos and generated thumbnails in ./data/storage on the host.

L35portsNetwork PortHost Network Port Binding
    ports:

Exposes the HTTP API and embedded web gallery to your local network or reverse proxy.

L368080:8080Network PortPort 8080 Mapping
      - "8080:8080"

Maps host port 8080 to container port 8080 (accessible at http://localhost:8080).

L37ai-serviceService RootOn-Device AI Vision Sidecar
  ai-service:

Isolated Python FastAPI microservice running InsightFace facial recognition, CLIP semantic search, RapidOCR, and offline reverse geocoding.

L38imageContainer ImagePre-Built AI Sidecar Image
    image: ghcr.io/nomidocloud/nomido-photo-ai-service:latest

Contains ONNX Runtime and pre-configured vision pipelines with zero cloud API calls.

L39restartLifecycleAI Service Restart Policy
    restart: unless-stopped

Keeps the local machine learning inference server running across reboots.

L40environmentEnvironmentAI Model & Runtime Environment
    environment:

Controls which InsightFace ONNX model tier is loaded into memory.

L41FACE_MODELHardware TunedInsightFace Detection Model (buffalo_m)
      FACE_MODEL: buffalo_m

Configured for Medium mode (4 GB RAM): uses buffalo_m (buffalo_s for Low, buffalo_m for Medium, buffalo_l for High) for optimal balance between RAM and facial clustering precision.

L42PYTHONUNBUFFEREDEnvironmentReal-Time Container Logging
      PYTHONUNBUFFERED: 1

Flushes Python stdout/stderr immediately so model loading and inference logs appear live in docker compose logs.

L1versionGlobal SpecCompose Specification Version
version: "3.8"

Declares Docker Compose file format 3.8, supporting service healthchecks, dependency conditions, and bridge networking.

L2servicesRoot TopologyContainer Services Root
services:

Defines the isolated multi-container stack: PostgreSQL vector database, Rust API backend, and Python AI sidecar.

Complete Sample .env File

# ==============================================================
# Nomido-Photo-Backend Configuration (.env)
# ==============================================================

# Database (PostgreSQL with pgvector)
DATABASE_URL=postgres://photo_user:secure_password_here@db:5432/photo_db

# Security & Authentication (Generate with: openssl rand -hex 32)
JWT_SECRET=please_generate_a_random_secret_at_least_32_chars_long

# Storage & Paths
STORAGE_PATH=./data/storage
LOG_PATH=./data/logs
LOG_LEVEL=info

# Network & Server Port
SERVER_PORT=8080
APP_URL=https://photos.yourdomain.com

# AI Microservice (Sidecar)
AI_SERVICE_URL=http://ai-service:8000
FACE_MODEL=buffalo_m
TOTAL_THREAD_NUMBER=4

# --------------------------------------------------------------
# Local-Only / Air-Gapped Mode:
# Leave HUB_API_KEY empty to disable all external cloud relay.
# --------------------------------------------------------------
# HUB_API_KEY=

Custom Reverse Proxy Setup (Nginx Example)

server {
    listen 443 ssl http2;
    server_name photos.yourdomain.com;

    ssl_certificate /etc/letsencrypt/live/photos.yourdomain.com/fullchain.pem;
    ssl_certificate_key /etc/letsencrypt/live/photos.yourdomain.com/privkey.pem;

    client_max_body_size 500M; # Support large camera video uploads

    location / {
        proxy_pass http://127.0.0.1:8080;
        proxy_http_version 1.1;
        proxy_set_header Upgrade $http_upgrade;
        proxy_set_header Connection "upgrade";
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header X-Forwarded-Proto $scheme;
    }
}