YouTube Language Level

YouTube Language Level

CEFR level checker for YouTube videos

Analyze the CEFR level (A1–C2) of any YouTube video using AI — cloud (Gemini API) or fully local, no internet required (Gemini Nano, Ollama).

Choose between three AI engines: Gemini API (cloud, free tier), Gemini Nano (built into Chrome), or Ollama (local server). Works for any language.


Available on Chrome Web Store
☕ Support on Ko-fi

This extension relies on free external AI services to analyze language level. These services may become paid or unavailable in the future — if you'd like to help keep the extension running, consider supporting it. Thank you! 💙

🌐 Language

Screenshots

CEFR level badges on YouTube thumbnails

CEFR level badges shown on YouTube video thumbnails

Extension popup — Gemini Nano tab

Extension popup — Gemini Nano tab

Extension popup — Ollama tab

Extension popup — Ollama tab

Extension popup — Gemini API tab

Extension popup — Gemini API tab

Features

🏷️ CEFR Badges

Colored circles (A1–C2) on YouTube video thumbnails

🤖 Three AI Engines

Use Gemini API (cloud), Gemini Nano (built-in Chrome AI), or Ollama (local models)

🌐 Multi-language

Analyzes videos in English, Spanish, French, German, Japanese, and more

🔒 100% Private

Everything runs locally — no data leaves your machine

🎸 Custom Server

Point to any Ollama instance on your network

⚡ Fast Cache

Results are cached locally to avoid re-analysis

📋 Adjustable Sample Size

Choose how many characters of the transcript to analyze (3000/6000/12000) to trade off speed vs. accuracy

Requirements

Option 1: Gemini API

Gemini API uses Google's cloud AI models with a free API key from Google AI Studio. Works on any Chromium-based browser and doesn't require downloading any model.

The Gemini API has a daily quota (resets at midnight Pacific time). Once you use it up, you'll need to wait for it to recharge before using it again.

1. Get your API key

  1. Go to Google AI Studio: aistudio.google.com/apikey
  2. Sign in with your Google account
  3. Click "Create API key"
  4. Copy the generated API key, paste it into the extension's API key field, then click OK
The Gemini API has a free tier with generous usage limits — no credit card required to get started.
Never share your API key with anyone — treat it like a password.

2. Configure in the extension

  1. Select the API Gemini tab
  2. Paste your API key into the field
  3. Click OK to save and test the key
  4. Select a model from the dropdown

3. Reload YouTube

Reload any open YouTube tabs so the extension can start working

Option 2: Gemini Nano

Gemini Nano is Chrome's built-in AI model. You need to download the AI model first.

Chrome is recommended for Gemini Nano. It may not work in other browsers.
Doesn't work in your browser? Use the Ollama option below instead — it works on any Chromium-based browser.
A Gemini Nano model will be downloaded. Do not close the browser until it is ready.

1. Activate Nano AI

  1. Enter this in the browser address bar:
    chrome://flags/#prompt-api-for-gemini-nano
  2. Set the flag to "Enabled Multilanguage"
  3. Click "Relaunch" or restart the browser
If the model does not start downloading, also enable (recommended):
chrome://flags/#optimization-guide-on-device-model and select "Enabled BypassPerfRequirement"

2. Check the model status

Open the YouTube Language Level popup and select the Gemini Nano tab:

StatusMeaning
AvailableReady to use
DownloadingModel is being downloaded
DownloadableNeeds to download first
UnavailableNot supported in your browser or model not downloaded

3. Choose the analysis language

Select the language of the video you want to analyze:

CodeLanguage
enEnglish
esSpanish
jaJapanese
deGerman
frFrench
Gemini Nano supports multilingual analysis. Select the language that matches the video content.

4. Reload YouTube

Reload any open YouTube tabs so the extension can start working

Option 3: Ollama

Works on any Chromium-based browser: Chrome, Brave, Edge, Vivaldi, Opera, and more.

1. Install Ollama

Linux / macOS:

curl -fsSL https://ollama.com/install.sh | sh

Windows:

Download the installer from ollama.com/download and run it.

2. Download a model

Run this in a terminal (Linux/macOS) or PowerShell/CMD (Windows):

ollama pull gemma3:1b
You can use any model from the Ollama model library — select it from the Ollama tab in the extension popup. A lightweight/small model (like gemma3:1b) is recommended for faster responses.

3. Configure CORS

The extension needs permission to talk to Ollama from YouTube.

Linux — Systemd (permanent)

sudo mkdir -p /etc/systemd/system/ollama.service.d
echo '[Service]
Environment=OLLAMA_ORIGINS=*' | sudo tee /etc/systemd/system/ollama.service.d/override.conf
sudo systemctl daemon-reload
sudo systemctl restart ollama

Linux — Temporary

sudo systemctl stop ollama
OLLAMA_ORIGINS=* ollama serve

Windows — Permanent

  1. Open System Properties → Environment Variables
  2. Add a new System variable: OLLAMA_ORIGINS = *
  3. Click OK and restart Ollama

Windows — Temporary (PowerShell)

$env:OLLAMA_ORIGINS="*"
ollama serve
If you want to use Ollama from another PC on your network, open Ollama Settings and enable "Expose Ollama to network". This allows connections from other devices on your local network.

4. Configure in the extension

  1. Click the extension icon
  2. Select the Ollama tab
  3. Set your server URL (default: http://localhost:11434)
  4. Click OK to test the connection
  5. Select a model from the dropdown

5. Reload YouTube

Reload any open YouTube tabs so the extension can start working

Using the Extension

  1. Go to https://www.youtube.com
  2. Videos with transcripts show a green spinner while analyzing
  3. A colored circle appears with the level: A1, A2, B1, B2, C1, or C2
  4. Hover over the badge to see which engine and model was used
  5. Click the extension icon to open the popup and switch between engines

Custom Ollama Server

By default the extension connects to http://localhost:11434. To change it:

  1. Open the extension popup
  2. Select the Ollama tab
  3. Enter your server URL (e.g. http://localhost:11434)
  4. Click OK — the extension will test the connection and load available models