Your Own AI Model (LM Studio, Ollama, Gemini Key), Step by Step

If you have an AI model on your own computer or an API key, JayNest can ask it for an item's name, category and properties. You enter the address of an OpenAI-compatible server, a key and a model. Your photos then go to your server, never to JayNest.

On iPhone

  1. In the Settings tab, tap Recognition, then under Recognizes the Item tap Your Own AI Model. The server settings screen opens.
    Your own AI model: address, key and model
  2. In the address field, enter the server address:
    • a computer at home: its address on your home network with the port, such as http://192.168.1.20:1234 for LM Studio or http://192.168.1.20:11434 for Ollama;
    • Gemini: https://generativelanguage.googleapis.com/v1beta/openai.
  3. If the server needs a key, enter it in API Key (Optional). For Gemini, it's a key from Google AI Studio.
  4. Tap Test Connection. For a computer at home, your iPhone asks for access to the local network: allow it. Once connected, you see how many models the server has, and the Model field turns into a list. Choose a model.
  5. Go back to the Recognition screen: Your Own AI Model has a checkmark, and its row shows the server address.
  6. Give it a try: take a photo in a new item. The note under the photos will say that the suggestions come from your model.

Server on Your Computer

Good to Know