How live translated captions work in a video call
A translated caption depends on two steps. Speech recognition turns audio into written words. Translation uses those words to produce text in the other person’s reading language.
See Hanashi caption optionsCompare the original and the translation
On Hanashi you can follow original text alongside translated captions. If a name is already wrong in the original caption, the translation may be carrying that mistake forward. If the original is right but the translated line sounds odd, the translation may have chosen the wrong meaning.
For example, if the original is “we’ll meet at two” when the speaker said “we’ll meet at the zoo,” ask them to repeat the destination. If the original has the right word with two meanings, ask for a more specific sentence.
Allow time for the words to arrive
A speaker’s pause gives the system a clearer boundary for the sentence. Give the other person time to read before continuing. When a word or time matters, type it or confirm it aloud.
Speech processing and transcript handling depend on the call’s applicable policies. See Hanashi’s privacy policy before using captions for a conversation you consider sensitive.