Skip to main content

AI & Translation

Open Sub Accounts → select a location → AI & Translation tab.

Translation is usually the first reason teams open this tab: each sub-account can translate inbound messages, outbound messages, or both, with explicit languages. The same area holds AI provider and model settings for features that depend on them. Everything is scoped to one GHL location (sub-account).

Where to configure​

  1. Open the AI & Translation tab on the sub-account.
  2. Enable inbound and/or outbound translation and pick languages (source/target as shown in the UI).
  3. Optionally enable use the contact’s language on outbound when the platform knows it.
  4. Configure AI provider, model, and any API keys or credentials the UI requests. Missing keys usually show a clear warning state.

Never paste secrets into chat or email; use only the fields provided in the app.

Setup and how it works​

Configure the AI provider, API key, and model; enable inbound translation, outbound translation, contact language, and speech-to-text; then verify the result in GoHighLevel and WhatsApp.

Forcing translation on one outbound (#traslate)​

In the GoHighLevel composer you can add #traslate or #traslate:xx (for example en, es) so that this outbound is translated before it reaches WhatsApp. The hashtag is removed in the delivered text.

Details and examples: Message codes from GoHighLevel.

Spintax and variations​

Some organizations use spintax (e.g. {hola | hi | hey}) in outbound copy composed in GHL. That is independent of translation but can be combined according to your workflow. Configure any spintax options your UI exposes on the same sub-account.

Good practices​

  • Align outbound language with what the contact expects; use contact-language mode when you serve multiple locales.
  • After rotating API keys, send a test message through a sandbox session if you have one.
  • Document for your team which conversations are auto-translated so agents are not surprised in the inbox.