Your AI is great with words,
we make it good with visuals.
LLMs, by nature, are just really fast next-word-predictors. This makes them well-suited for routine conversations, getting answers, checking grammar, etc. but leave a lot to be desired from tasks that need visual meaning.
Current AI-generated artefacts tend towards a "creative median" and start to all look the same — hence the term, AI slop. We see this as an opportunity to intervene.
We're currently working on a "harness" — a set of tools and constraints around an LLM — to make it better suited for tasks that need visual ingenuity or design awareness.
We hope to do for design and media what Claude Code and Devin did for software development, and make better design accessible to more people as a result.
Our past work in the space →
We work with creative teams to scale their design and media.
We build and improve our platform via design partnerships, often by consulting companies in the following situations:
- They have a sophisticated brand language that they want to create a content generation system around.
- They have an inventory of products that need some creative storytelling in order to sell/market better.
- They have a design collateral that is operationally challenging to create that they'd like to automate.
- They are building a product within which they'd like to offer generative design/media via Helmer APIs.
We're a small, passionate team, inventing a more creative future.
We care about the messy middle between taste and infrastructure: the systems that turn ambiguous creative processes into tools people can actually direct.
Our team has experience across design tools, video infrastructure, AI agents, and creative automation, and we're drawn to hard creative problems where better tools can change what people are able to imagine and ship.
Come join us — we're hiring!
Backed by
