Appearance
AI Models
LingoLark can fill English glosses with AI, or leave the English side blank for you to write. Pick the path that matches how much setup you want and whether prompts may leave your machine.
Model options
LingoLark Cloud
Use LingoLark Cloud when you do not want to install a local model. You sign in, spend one AI credit per successful action, and LingoLark runs the prompt on its servers. Alpha: testers need a granted credit balance; buying packs in the app is not live yet.
Ollama
Use Ollama when you want the most predictable local setup. You install Ollama on your machine, pull the model LingoLark expects, and the extension calls the local server at http://127.0.0.1:11434.
Chrome Nano
Use Chrome Nano when Chrome already has Gemini Nano available on your device. This route is browser-managed and does not use Ollama, but today it depends on Chrome already having the model ready.
Which should I choose?
| If you want... | Start with |
|---|---|
| No local model; sign in and use credits | LingoLark Cloud |
| A setup you can install and verify yourself | Ollama |
| Chrome-managed on-device AI with little extension setup | Chrome Nano |
| No automatic AI glosses | Type it yourself |
You can switch translation models from the LingoLark toolbar popup. In the web Reader, a Cloud / Local switch in the language tools rail does the same job for those tools.
Alpha Cloud notes
- Cloud needs a signed-in session and a remaining credit balance.
- One successful Cloud action = one credit. A multi-word extension save is one Cloud call.
- Out of credits, or not signed in: use Local / Nano / Type it yourself. Those paths stay free.
- Cloud prompts go to OpenRouter and the upstream model. A free-tier model may train on prompts -- see Privacy Policy 4.e and the Cloud page.