Foundry’s Griptape looks to bring structure and predictability to AI workflows

Foundry’s Martin Mayer discusses how Griptape aims to orchestrate multiple AI agents, introduce repeatable workflows and bridge the gap between generative AI experimentation and practical VFX production.

As artificial intelligence continues to reshape creative workflows, the challenge for visual effects studios is increasingly moving beyond what AI can generate to how these capabilities can be integrated into predictable, production-ready pipelines. Foundry, the company behind industry tools such as Nuke, Katana and Mari, is exploring this space with Griptape, its AI orchestration tool.

Speaking about the company’s latest offering, Foundry senior manager, creative services Martin Mayer, explained that Griptape is designed to coordinate AI-driven tasks across multiple agents and stages of a workflow.

“We are the makers of Nuke at Foundry. We made Nuke, which is a software for visual effects, Nuke, Katana, Mari and the most recent addition Griptape, is an AI orchestration tool,” Mayer said.

Orchestrating multiple AI agents

Griptape enables users to build workflows involving different AI models and agents, including large language models (LLMs) and image-generation models. Rather than treating each AI interaction as an isolated task, the platform allows users to coordinate multiple steps into a structured process.

“It allows you to orchestrate AI work across multiple agents and steps,” Mayer explained. “Whether the steps or agents are based on text, which would be the LLMs, or the image-generation models, you can really create a predictable, repeatable workflow across all of this.”

This approach is intended to make AI workflows more manageable for creative professionals, particularly in environments where multiple artists, supervisors and technical teams need to work together.

One of the ideas behind Griptape is to move beyond conventional text-based prompting towards a more visual approach to communicating creative intent. Instead of relying entirely on written instructions, users can define what an AI agent should see, understand and create, helping establish greater consistency in the results.

Mayer described this as a shift towards “visual prompting”, with the ability to define and capture creative instructions in a way that can be repeated across workflows.

Bringing order to the AI boom

For Mayer, one of the biggest challenges with the current wave of AI adoption is the largely organic way in which tools and workflows are developing. While experimentation is accelerating, this does not always translate into the productivity and predictability required for professional VFX production.

“There are many goals, but one thing that I feel, especially in the whole AI segment, is that it’s just growing very organically. And that doesn’t necessarily align with the productivity, predictability that we need to actually create successful shots and communicate the changes across many, many people,” he observed.

He believes Griptape could help address this gap by acting as a common framework for communicating between AI agents, artists and the people developing creative workflows.

“Griptape could be potentially the language,” Mayer said. “It brings in a way of communicating, not only with the agents and the AI, but also with people who develop and make these workflows and work with AI.”

He added that the platform could be viewed as “a communication and potential standardisation layer, among many other things.”

This is particularly relevant to VFX pipelines, where creative decisions, revisions and technical requirements need to be communicated across several departments. A structured approach to AI orchestration could help make these processes more consistent and easier to manage.

Moving from hype to practical applications

While acknowledging the transformative potential of AI, Mayer believes the industry is still navigating a period of considerable experimentation and hype. The next step, according to him, is to identify tools and approaches that can deliver tangible value in everyday production.

“I think we are going through a period of huge hype and it needs to settle down to something we can actually practically use.” Mayer said.

He believes the industry is beginning to see early examples of that transition, with tools such as Griptape offering a way to introduce structure into an otherwise fragmented AI ecosystem.

“We are seeing the first glimpses of it. Griptape could be viewed as one of them, bringing order into the chaos of the AI boom, at least some segment of it,” Mayer said.

As AI capabilities continue to expand, Foundry‘s approach through Griptape reflects an important shift in the conversation around creative technology: from simply generating content to building reliable systems through which artists and AI tools can work together. For the VFX industry, that distinction could prove crucial in making AI a practical part of production rather than just another experimental tool.

Follow us on Google News