Luma unveils unified AI agents to streamline creativity

Luma has launched Luma Agents, a set of AI tools for creative teams that work across text, images, video and audio within a single system.

The product is positioned as an AI collaborator that maintains context across a project, from the initial brief through delivery. Luma says it can manage iterations and coordinate different tools and models during production.

Amit Jain, Luma's co-founder and CEO, said the product is a response to capacity constraints in creative teams.

"Creative work has never lacked ambition, it's lacked execution capacity," Jain said. "Creative teams shouldn't have to spend their time orchestrating tools. They should spend it creating. Agents aren't shortcuts. They're collaborators that maintain context, coordinate execution, and advance projects so teams can focus on taste, direction, and strategy."

Agency deployments

Luma Agents are rolling out with enterprise partners, including Publicis Groupe Middle East and Serviceplan Group. The deployments cover strategy, creative development and production workflows, Luma says.

Serviceplan Group says it has integrated Luma into its internal AI programme and connected it to day-to-day creative processes across its network.

"Luma is now part of our broader House of AI ecosystem and integrated directly into our creative workflows. It allows our teams across more than 20 countries to collaborate more smoothly and develop great work faster. For our clients, that means high-quality creative output delivered with greater speed and efficiency - without compromising craft," said Alexander Schill, global CCO at Serviceplan Group.

Workflow model

Luma describes Luma Agents as an alternative to AI production workflows that rely on separate models for language, vision, video and reasoning. Those approaches often require orchestration layers that pass work between tools and reconstruct context at each stage.

Luma's approach aims to keep context intact as projects move between formats. In practice, Luma Agents operate in a shared environment where humans direct the creative work while the system handles routing and execution across tasks.

Luma says the product can run multiple creative directions in parallel and refine outputs through evaluation rather than relying on single-shot generation. It also integrates with enterprise tools and production systems via an API.

Unified Intelligence

Luma Agents run on what the company calls a Unified Intelligence architecture, which it says trains a single multimodal reasoning system to understand and generate across formats within the same model design.

The first model in this line is Uni-1. Luma describes Uni-1 as a decoder-only autoregressive transformer that uses a shared token space for language and image tokens, enabling it to reason in language while generating images in a single forward pass.

Luma says the design links reasoning and rendering more closely than pipeline approaches that hand work from one specialist model to another. It says this enables planning and production to occur as a single process, rather than as separate steps managed by an external workflow layer.

Alongside its own model, Luma says Luma Agents can coordinate work across third-party models, including Ray3.14, Veo 3, Sora 2, Kling 2.6, Nano Banana Pro, Seedream, GPT Image 1.5, and ElevenLabs. The system can select and route tasks to different models based on the required step, the company says.

Jain said the company's strategy is to treat multimodal work as a unified problem rather than separate domains.

"Intelligence shouldn't be fragmented by modality," Jain said. "Unified systems reason holistically. When the same model can think, imagine, and render, you move closer to intelligence that behaves coherently across the entire creative process."

Enterprise controls

Luma says Luma Agents include controls for enterprise customers, such as customer IP ownership, automated content review to reduce copyright risk, documentation intended to show human involvement, required human-review workflows before public release, and cloud-based infrastructure with enterprise guardrails.

Luma says further deployments with global enterprise partners are underway as agencies and brand teams assess how AI systems fit into production processes spanning multiple formats and markets.