Platforms

all

Description

We'd like to improve visibility into a common issue users are experiencing with Leo language mismatches between prompt language and model response language. The work to resolve language detection already exists server side, so what we'd like to do is include the user's "selected_language" value when they submit feedback about the conversation from the client.

• Add "selected_language" param to /1/ai/feedback/rating request payload
• Include a new category in the feedback dropdown for "Language Issue" (or something similar)

On the reporting side some backend work is required detailed here: https://github.com/brave-intl/go-vault/issues/247