Batch Processing/Local Processing

Local Processing

MetaPhotoAI offers the ability to generate image metadata using local AI models. This allows you to process your photo library on your own hardware without using cloud credits or requiring an internet connection.

Warning

While photos can be processed locally, video metadata generation is currently only available online. All video analysis requires an active internet connection and will consume your monthly cloud credits.

Benefits of Local Photo Processing#

BenefitDescription
No Credits RequiredMetadata generation for images is free and does not consume your monthly credit allowance
PrivacyAll image analysis is performed on your machine; no data ever leaves your computer
Offline AccessProcess your photo library anywhere, even without a network connection

Prerequisites for Local AI#

Local AI requires significant computing power. You will need:

  1. A computer with sufficient CPU/GPU resources and RAM
  2. Ollama: A third-party application used to run large language models locally

Setting Up Ollama#

Check status

In MetaPhotoAI, go to Settings > Local Processing.

Download

If Ollama is not installed, click the link provided in the instructions or visit ollama.com/download.

Install

Run the installer for your operating system (macOS, Windows, or Linux).

Launch

Open the Ollama application. You must ensure the Ollama server is running for MetaPhotoAI to connect to it.

Downloading AI Models#

Before you can process photos locally, you must download a compatible AI model:

ModelRecommendation
Gemma 3 4BRecommended for mid-range hardware
Gemma 3 12BProvides higher accuracy but requires more system resources

Download Command#

Open your computer's Terminal (macOS) or Command Prompt (Windows) and type:

ollama pull gemma3:4b

For the larger model:

ollama pull gemma3:12b

Note

These files are several gigabytes in size and may take time depending on your internet speed.

Enabling and Using Local Processing#

Enable in Settings

Go to MetaPhotoAI Settings > Local Processing and toggle "Enable Local Processing" to ON.

Check the Local toggle

In the top navigation bar, you will see a "Local" button.

Process media

Select your images and click "Generate Metadata." A blue "LOCAL" badge will appear next to the processing task in the right panel.

Local Toggle States#

StateMeaning
GreenLocal processing is active (Photos only)
GrayCloud processing is active (Uses credits for photos and videos)

Performance Expectations#

Local processing speed is entirely dependent on your computer's hardware:

Processing TypeTypical Speed
Cloud ProcessingHighly optimized and fast (e.g., 2 images in ~3 seconds)
Local ProcessingSlower, depending on your CPU/GPU (e.g., 2 images in ~25 seconds)

System Requirements#

For optimal local processing performance:

ComponentRecommendation
RAM16GB minimum, 32GB recommended
GPUDedicated GPU with 8GB+ VRAM recommended
StorageSSD with 10GB+ free space for models
CPUModern multi-core processor

Tip

By setting up local processing, you gain a private and cost-effective way to manage your photo metadata, while reserving your cloud credits specifically for video content.

Troubleshooting#

IssueSolution
Ollama not detectedEnsure Ollama is running in the background
Slow processingTry the smaller 4B model or upgrade hardware
Connection errorsRestart Ollama and MetaPhotoAI
Model not foundRun the ollama pull command again

Next Steps#