1 · The Product
What is LOFT?
The OS for AI Production - an Integrated Production Environment (IPE) for generative AI. It connects to whatever models a studio chooses and provides the durable structure beneath them: assets, lineage, provenance, review, cost, and a shared intent-to-action record.
What isn’t it?
Not a model, a generation engine, or a marketplace. LOFT doesn’t make pixels - it makes them accountable. It turns model output into traceable, production-ready assets.
Why call it an operating system?
An OS isn’t the applications; it’s the layer that makes them usable - owning resources, abstracting devices, enforcing permissions, keeping durable state. Models are the applications here, and they churn. Every studio adopting AI is accumulating dependencies that will be gone or repriced within two years. The studios that survive that will be the ones whose assets and decisions live in a layer that doesn’t care which model made the pixels.
How is this different from Kling, Higgsfield, Luma, or Runway?
Those are applications optimized for creation. LOFT is optimized for production readiness. They help teams make content faster; LOFT helps organizations manage, govern, and scale it. The competition shifts from model novelty to workflow depth - harder to replicate, and more durable.
Why does model neutrality matter?
Because the model layer keeps churning. LOFT treats models as interchangeable services around a stable production core, so the production record survives that churn - and becomes more valuable as it accelerates.
2 · Capabilities
What does LOFT do?
It combines four things in one system: a system of record for assets, versions, lineage, and decisions; multi-model orchestration through a consistent interface; integrated review, approval, cost tracking, and governance; and editorial, where the cut is a first-class asset referencing its source material.
What’s the workflow?
Generate → Track → Review → Commit → Playbook → Govern → Deliver. Successful combinations of prompts, models, and settings can be saved as reusable playbooks.
What makes editorial different?
Most pipelines lose provenance the moment generated material enters an edit. LOFT’s cut references its takes rather than copying them - swap a take and the cut re-points instead of forcing a manual reconform.
Does it replace the DCC tools artists use?
No. It sits alongside them, managing the assets, versions, and AI activity around the work while artists stay in their established applications.
What hardware is required?
Generally none. Hosted providers run inference on their own infrastructure, and bring-your-own-key means credentials never pass through LOFT’s servers. Storage and bandwidth still matter for large assets.
3 · Proof
Has it shipped?
Yes. LOFT is production-ready and shipping to select customers through guided onboarding.
What has it been proven on?
Two eras. First, LOFT was built as a resource and production management tool with Jellyfish Pictures - deployed studio-wide to hundreds of users, it transformed their business model with measurable efficiencies in production and finance. Today, it is shaped by independent collaboration with Phil Tippett on AI-driven productions including Sentinel, Pequin, and the Mad God 2 (Mad Goddess) teaser.
What is Phil Tippett’s involvement?
Advisor to LOFT. Two Academy Awards; the creature and animation work behind Star Wars, RoboCop, Jurassic Park, and Starship Troopers; director of Mad God. He is not a figurehead - the requirements that became LOFT were written alongside his films.
Is LOFT affiliated with Tippett Studio?
No. LOFT and Tippett Studio are unaffiliated. LOFT was developed 100% outside of any studio, conceived and built independently, and available to any studio. Both founders previously worked with Tippett Studio and continue to collaborate with Phil Tippett on independent productions.
4 · Company
Who founded it, and when?
LOFT was founded in 2022 in Berkeley, California by Gary Mundell, CEO, and Remko Noteboom, CTO.
Why are these founders credible?
They’ve built production infrastructure before. Both were on the original Alias|Wavefront Maya development team. Noteboom built Nelnet at Nelvana - the world’s first digital production workflow tool - then developed TACTIC, deployed across 350+ studios worldwide. The Nelnet work earned a 2021 Technology & Engineering Emmy Award. More than 25 years each, through every major production shift.
How does pricing work?
Seats buy authorship; credits meter generation. Free seats let reviewers and producers participate at no cost. Customers can also bring their own model accounts and pay providers directly, with LOFT charging only for seats.
Where is the company based?
Berkeley, California