NextAI Translator

Translate, polish, and summarize with TokenMix models.

Prepare TokenMix values

Provider: OpenAI
API Key: your TokenMix key
API URL / API Host / Proxy URL: https://api.tokenmix.ai
API URL Path: /v1/chat/completions
Model: a TokenMix model ID

Browser extension setup

Install the NextAI Translator browser extension.
Click the extension icon and open settings.
Choose OpenAI as provider.
Paste your TokenMix API Key.
Set API URL or Proxy URL to https://api.tokenmix.ai.
If there is an API URL Path field, set /v1/chat/completions.
Choose custom model and enter a TokenMix model ID.
Save settings and refresh the page.

Desktop setup

The desktop app uses the same idea in Settings. For Windows/macOS/Linux permission prompts, follow the official README. TokenMix integration still depends on Key, API URL, Path, and Model.

Test

Select a short paragraph on a web page and run translate. Recommended test prompt:

Translate the selected text into natural English. Keep the meaning accurate and avoid adding extra information.

Troubleshooting

request path contains /v1/v1: API URL should be https://api.tokenmix.ai, not /v1.
401: wrong API key.
404/model not found: wrong model ID.
no response: save settings, refresh the page, or restart the extension/app.
poor style: try another TokenMix model or adjust the custom action prompt.

Practical notes

NextAI Translator should clearly separate API URL and API URL Path. The source joins https://api.openai.com with /v1/chat/completions, so TokenMix should be split the same way.

Recommended config

Provider: OpenAI
API URL: https://api.tokenmix.ai
API URL Path: /v1/chat/completions
API Key: your TokenMix key
Model: a TokenMix text model ID

Beginner test order

Translate a short selected paragraph on a web page.
Then test summarize.
Finally tune polish, custom actions, vocabulary, or other features.

If users enable many custom actions first, it becomes hard to tell whether the issue is TokenMix connectivity or prompt/action configuration.