AI audio production platform AudioStack has added Gradium’s voice models to its platform, the companies said on August 19, 2026. The partnership secures new distribution for Gradium and enhances AudioStack’s offering for media and advertising customers, helping the two companies differentiate themselves in the increasingly crowdedvoice AI market.
The integration places Gradium models inside AudioStack’s existing workflow, making them available across multiple languages and regional accents. Customers can select Gradium voices through AudioStack’s API or console and combine them with script generation, music, sound design, mixing, and mastering.
According to the partnership announcement, AudioStack selected Gradium partly for its coverage of regional speech. The companies cited Australian and French Canadian voices, as well as Dublin English and Bavarian German, among the available options.
Partner spotlight
Fast isn’t good enough.
Download Checklist
Partner spotlight
Riding the AI shockwave
The future of AI adoption depends on trust, ethics and human expertise.
“We were impressed by Gradium’s range and consistency, offering high-quality regional accents, such as Australian, alongside French-language depth that includes French Canadian voices, which have already helped AudioStack support customers in that market,” said AudioStack CEO Alexis van de Wyer.
Gradium CEO Neil Zeghidour positioned the integration as a way for advertisers to insert regionally appropriate voices directly into production pipelines.
AudioStack Expands Its Provider Layer
AudioStack presents itself as an orchestration and production platform rather than an AI voice developer. Its system aggregates models from external providers and handles the additional steps required to turn generated speech into completed audio assets. Founded in 2019, AudioStack disclosed a GBP 1m seed round in 2021 and a USD 3.1m pre-Series A round in 2024.
In a May 2025 overview, AudioStack said its catalog contained more than 1,700 voices from over a dozen providers. The roster at that time included ElevenLabs, Cartesia, OpenAI, Microsoft Azure, Resemble.ai, WellSaid Labs, and several cloud text-to-speech services.
The multi-provider approach lets AudioStack route production work among different models based on language, accent, style, or use case. It also puts the company between voice developers and end customers, with responsibility for provider integration, pronunciation controls, and production standardization.
That infrastructure layer was a focus of SlatorCon London in May, 2026. During a panel featuring Zeghidour and AudioStack CTO and co-founder Peadar Coyle, participants discussed model costs, workflow integration, quality assurance, governance, and support for regional languages. Coyle said voice models alone were insufficient for production deployments.
Gradium Builds Distribution After Funding Extension
The AudioStack deal follows Gradium’s July 2026 announcement that it had increased its total seed funding to USD 100m. NVIDIA joined the extension alongside existing investors including FirstMark Capital, Eurazeo, and DST Global Partners. Gradium had previously announced a USD 70m seed round in December 2025.
Gradium was founded in September 2025 by researchers associated with French AI lab Kyutai. Its products cover text-to-speech, speech recognition, voice cloning, live translation, and developer infrastructure for conversational AI.
As funding and new entrants intensify competition among voice AI providers, Gradium will need to differentiate its models while securing distribution through production platforms such as AudioStack.
