ElevenLabs Ships Dubbing v2, Music v2, and Ads Engine Across Voice AI Platform.

ElevenLabs rolled out a wave of updates across its voice and audio platform in June, including improvements to dubbing and music generation, new agent workflows, and a localization engine for marketers.

The company announced the releases in a LinkedIn post recapping a month that also included crossing $500 million in annual recurring revenue, partnerships with four governments, and an investment from Poland.

Dubbing v2 aims to preserve the emotion, performance, and intent of original content during localization. The company described the update as addressing flat, disconnected dubbing, though it did not provide technical details on how the new model achieves emotional fidelity across languages.

Music v2 delivers improved vocals, instrumentation, arrangement, and multilingual support across genres. The upgrade follows the platform's earlier music generation capabilities, which ElevenLabs has positioned for creators and audio producers.

ElevenLabs introduced three features aimed at teams building conversational AI:

Procedures package sets of instructions that define how agents handle common scenarios, similar to standard operating procedures for employees. The feature, recently covered, allows developers to codify workflows like returns, appointment scheduling, or escalation paths.

New modalities enable agents to process images, files, audio notes, contacts, and location pins within customer conversations. The expanded input types let agents handle richer interactions beyond text and speech.

Speech Engine converts existing chat agents into voice agents using a single prompt. The tool is designed to simplify the integration of voice capabilities into existing conversational systems.

The Ads Engine automates ad localization across more than 50 languages and pushes refreshed creative back into ad platforms. The feature targets marketers running multilingual campaigns who need to maintain brand voice and messaging consistency across regions. It sits within ElevenCreative, the company's suite for content production.

Avatars create talking-head videos from a script, voice, and photo. The tool generates synthetic presenters for video content without requiring filming or on-camera talent.

Flows Agent lets teams build creative workflows through natural language in a node-based canvas. The interface aims to let non-technical users assemble multi-step audio and video production pipelines by describing steps conversationally rather than writing code.

ElevenLabs also added watermarking techniques for content authenticity. The company embedded SynthID watermarks in AI-generated audio and launched a free detector tool in late June, working with Google DeepMind on the implementation.

The Voice Library, where users can license their voices for use in the platform, has paid out more than $22 million to contributors. The company said it has more product launches planned for the third quarter but did not detail what those releases will include.

ElevenLabs hosted a summit in Warsaw with 2,500 attendees in June and announced partnerships with the governments of Poland, the United Kingdom, California, and one additional government it did not name. The Poland deal included an investment from Vinci, part of the BGK Group, joining venture firms Andreessen Horowitz, Sequoia, and ICONIQ as investors.