AI Tagging

You need to set up an AI provider to use the AI tagging feature in your self-hosted instance of the app. This feature allows you to automatically tag your media files with relevant keywords, making it easier to search and organize your content.

Supported Providers

Ollama Provider

You can set up Ollama and pull your preferred model onto your server to use this feature. A lightweight model like the phi3:mini-4k is enough for this feature.

After that, you need to define the NEXT_PUBLIC_OLLAMA_ENDPOINT_URL and OLLAMA_MODEL environment variables inside your .env file.

Here's an example:

NEXT_PUBLIC_OLLAMA_ENDPOINT_URL=http://localhost:11434
OLLAMA_MODEL=phi3:mini-4k

OpenAI Compatible Provider

You can use an OpenAI-compatible provider for AI tagging by defining the following environment variables in your .env file:

CUSTOM_OPENAI_BASE_URL=
OPENAI_MODEL=
OPENAI_API_KEY=

Note that if you want to use OpenAI's official API, you don't need to define the CUSTOM_OPENAI_BASE_URL variable, as it defaults to OpenAI's API URL. Here's an example of how to set it up:

OPENAI_MODEL=gpt-4o
OPENAI_API_KEY=YOUR_OPENAI_API_KEY

Another example, to use Gemini:

CUSTOM_OPENAI_BASE_URL=https://generativelanguage.googleapis.com/v1beta
OPENAI_MODEL=gemini-2.5-flash-preview-04-17
OPENAI_API_KEY=YOUR_GOOGLE_AI_STUDIO_API_KEY