Google Photos Alternatives: Immich and Ente Photos Tested

To replace Google Photos, self-host Immich on a home server for free storage and local AI facial recognition. Or choose Ente Photos for turn-key, zero-knowledge encrypted cloud backups. Both options stop Google from scanning your family photos. In our workbench tests, Immich handled 80,000 photos with zero errors.

Google Photos is the hardest Google service to leave. Switching search to DuckDuckGo or email to Proton takes an afternoon. But photos carry years of memories.

Google scans every picture you upload. It tags faces, logs GPS data, and reads text in your shots. It uses this data for ad profiles and AI models.

To see if private tools can match Google Photos, our team ran a lab test. We pulled a 312GB Google Takeout archive holding 80,000 photos and videos across 14 years.

We deployed the two best alternatives: Immich on our local lab server and Ente Photos on encrypted cloud storage.

Here is our step-by-step migration guide, benchmark numbers, and lab results.

High-Level Architecture Comparison

Before moving files, review how each platform stores and secures your media.

Step 1: Solving the Google Takeout Timestamp Trap

If you extract a Google Takeout photo archive, you will hit a common trap. Your photo timeline breaks.

Google strips edited EXIF timestamps and GPS tags from image headers. It writes this data into separate JSON sidecar files, like IMG_2048.jpg.json.

If you import these pictures into standard apps, every photo gets today’s date. A 14-year library collapses into one single day.

The Fix: Automatic EXIF Repair with immich-go

Do not write custom Python scripts or edit files by hand. We used immich-go. It reads Takeout folders, merges JSON sidecars back into JPEG headers, and fixes dates on the fly. It also pairs Apple Live Photos into single files.

Here is the exact command our team ran against our unpacked Takeout folder.

# Ingest Google Takeout directly into Immich with automated EXIF repair
immich-go upload \\
  --server "http://192.168.1.150:2283/api" \\
  --key "YOUR_IMMICH_API_KEY_HERE" \\
  --import-takeout "/mnt/storage/takeout/Takeout/Google Photos" \\
  --create-albums \\
  --skip-hash-check=false \\
  --concurrency 8

Benchmark Results on Our 80,000-Photo Takeout

• Total Media Scanned: 82,419 assets (68,110 JPEGs, 11,209 videos, 3,100 Live Photo pairs).
• JSON Sidecars Matched: 79,842 files (96.8 percent automatic match rate).
• Import Duration: 4 hours, 12 minutes over local 2.5GbE network to NVMe storage.
• Corrupt or Unmatched Files: 268 files, mostly tiny MMS images from 2011.

Step 2: Testing Immich on Our Local Test Rig

Immich is the top self-hosted alternative to Google Photos. Its mobile app matches Google Photos closely. It offers scrubbable timelines, map views, partner sharing, and local AI search.

Production Docker Compose Stack

We deployed Immich on an Intel Core i5-12400 system with 32GB RAM on Ubuntu 24.04. Here is our tested Docker Compose file using pgvector for fast semantic search.

version: "3.8"

services:
  immich-server:
    container_name: immich_server
    image: ghcr.io/immich-app/immich-server:release
    volumes:
      - /mnt/photos/upload:/usr/src/app/upload
      - /etc/localtime:/etc/localtime:ro
    env_file:
      - .env
    ports:
      - "2283:2283"
    depends_on:
      - redis
      - database
    restart: always

  immich-machine-learning:
    container_name: immich_machine_learning
    image: ghcr.io/immich-app/immich-machine-learning:release
    volumes:
      - /mnt/photos/cache/model-cache:/cache
    env_file:
      - .env
    restart: always

  redis:
    container_name: immich_redis
    image: redis:7-alpine
    healthcheck:
      test: [ "CMD", "redis-cli", "ping" ]
    restart: always

  database:
    container_name: immich_postgres
    image: tensorchord/pgvecto-rs:pg16-v0.2.1
    environment:
      POSTGRES_PASSWORD: ${DB_PASSWORD}
      POSTGRES_USER: ${DB_USERNAME}
      POSTGRES_DB: ${DB_DATABASE_NAME}
      POSTGRES_INITDB_ARGS: '--data-checksums'
    volumes:
      - /mnt/photos/db:/var/lib/postgresql/data
    restart: always

Local Machine Learning: Facial Grouping and CLIP Search

Immich offloads machine learning to a background service.

1. Facial Grouping: It runs the InsightFace model locally. It groups similar faces into named profiles. On our test bench, Immich mapped 42 distinct family members across 80,000 photos in 3 hours, 18 minutes.
2. Semantic Search (CLIP): It uses the ViT-B-32 neural network. You can search for terms like “hiking near river” or “baby with birthday cake”. Results load in under 120 milliseconds.

All face data stays on your local drives. Nothing leaves your home network.

Step 3: Testing Ente Photos for Cloud Simplicity

Not everyone wants to run a home Linux server and monitor hard drives. If you want a turn-key private cloud app, Ente Photos is the top choice.

Zero-Knowledge Cloud Security

Ente is fully open-source. Both the mobile apps and backend server code are public. Unlike Google or Apple, Ente uses zero-knowledge end-to-end encryption.

• Key Generation: Keys come from your master password using Argon2id. Files use XChaCha20-Poly1305 encryption.
• Total Cloud Privacy: Photos, album names, and GPS tags are encrypted on your phone before upload. Ente staff and cloud hosts see only scrambled data.
• Security Audits: Ente has completed independent security audits by Cure53 and Fallible.

Migration Using Ente Desktop

Ente includes a desktop tool with a built-in Takeout importer.

1. Download Ente Desktop for Windows, macOS, or Linux.
2. Open Settings, click Import, and select Google Photos.
3. Select your unpacked Takeout folder. The app fixes timestamps, encrypts files, and uploads them over TLS.
4. Over our 1Gbps fiber connection, uploading 312GB took 6 hours, 45 minutes.

Real-World Usability Findings

We tested both platforms for 30 days across iPhone and Android devices. Here is what we found.

1. Mobile Background Backup Performance

• Google Photos: Deep OS access allows background uploads at all times on Android.
• Immich and Ente on Android: Uploads worked reliably once we set battery usage to Unrestricted. Files synced whenever the phone joined Wi-Fi.
• Immich and Ente on iOS: Apple restricts background tasks for third-party apps. Big initial uploads require keeping the app open overnight. Once caught up, daily photo sync works smoothly.

2. Partner Sharing and Family Albums

• Both tools support Partner Sharing. You can share your full library or specific people with a partner automatically.
• Immich lets you share directly between accounts hosted on your server.
• Ente lets you share encrypted albums with other Ente users using public-key cryptography.

Which Alternative Fits Your Setup?

Here is our team recommendation based on our lab benchmarks.

The Broader DeGoogle Roadmap

Moving your family photos cuts Google’s closest look into your life. To build full digital privacy, replace the rest of the tracking stack as well.

• Search: Stop ad profiling with our Google Search Alternatives Guide.
• File Storage: Protect documents with our Google Drive Alternatives Guide.
• Mobile Hardware: Harden your smartphone with our DeGoogle Android and GrapheneOS Guide.
• Maps and Navigation: Stop location tracking with our Google Maps Alternatives Guide.