Hudson AI Debuts Localization Agent at MIPCOM 2026
The agent, built on Hudson Studio, separates overlapping speech from more than 40 speakers and automates over 90% of quality control.
Hudson AI unveiled a localization AI agent that combines dubbing generation with quality control, presenting it for the first time at MIPCOM 2026 in Cannes, France, held October 12–15, according to Venturesquare. The agent runs on Hudson AI's dubbing platform, Hudson Studio.
The agent can identify more than 40 speakers in a video and separate overlapping speech. It automatically assigns character names and identifying information by analyzing the video's flow and the relationships between characters.
Hudson AI says the system automated more than 90% of quality control work, cutting the amount requiring human review to about one-tenth of previous levels. It also detects non-verbal sounds such as laughter, sighs and breathing and converts them to match the dubbed character's voice.
The company has completed dubbing quality tests for Japanese, English, Latin American Spanish, Brazilian Portuguese, Arabic and French. Following the event, it has been running proofs of concept with major global media companies using real video content.
Quick answers
Where was Hudson AI's localization agent first shown?
It was first presented at MIPCOM 2026 in Cannes, France, held October 12–15, 2026.
Which languages have been tested for dubbing quality?
Japanese, English, Latin American Spanish, Brazilian Portuguese, Arabic and French.
What platform is the agent built on?
It is built on Hudson AI's dubbing platform, Hudson Studio.