이 페이지의 한국어 번역은 준비 중입니다. 아래는 영문 원본입니다.
영문으로 보기 →FAQ
Five sections. Short answers. No marketing.
Jump to a section
01 / 05
$5,000 and up. Our entry-level productized sprints (AI Ops Audit, Web 3D Product Showcase) start there. Below that we can’t do our best work, and you’re better served by a capable freelancer. Happy to recommend one.
We’ve shipped projects up to around $300K solo-studio scope. Above that, we either pull in additional network collaborators or co-deliver with partner studios. If your project is $500K plus, we’ll be honest about whether we’re the right fit.
Yes. Our productized Brand + Site Express sprint ($14K, 4 to 6 weeks) was built specifically for funded startups that need a fast, opinionated brand-and-site build. We also take on custom engagements for AI-native startups and consumer-brand launches.
Yes, when the work fits our shape. We’re good at a well-scoped project inside a larger organization. A training app, a configurator, a product-launch site. We’re less good at multi-stakeholder enterprise procurement with long RFP cycles. Big agencies are better at that.
Yes. Retainers start at $4,000 per month for 20 hours of ongoing design, development, or 3D work. Standard contracts, cancel with 30 days’ notice. We don’t lock clients in.
02 / 05
Industrial explainers, operator training, assembly instructions, process visualizations, product configurators, and patient-education 3D. Our Interactive 3D Applications page has the full breakdown by project pattern.
Not usually. Our 3D work is technical. Machinery, products, processes. If you need character animation, in-game 3D, or film-VFX-tier work, we’ll refer you to a specialist.
Both. If you have CAD, we optimize it for browser delivery. If you don’t, we build the 3D asset from drawings, photography, or physical reference. It’s one of the advantages of being a full-pipeline studio. AI-assisted baseline geometry (Tripo, Hunyuan3D, Meshy) compresses the modeling stage significantly.
Three.js / React Three Fiber for most projects. Verge3D for industrial training where the client’s design pipeline is Blender-first. Babylon.js for simulation-heavy scenarios with particle effects and state. We avoid Unity WebGL for browser delivery because load weights are punitive.
Not native. We build mobile-responsive web apps that work in-browser on any device. If you need true native iOS / Android, we’ll refer you to a native-specialist studio.
Yes. We prefer Next.js plus Sanity for greenfield projects, but we ship in Webflow, WordPress (Elementor), Framer, Astro, and whatever else a client’s stack requires. CABEON was built in WordPress; enParticle and ROBOne as well. We adapt.
Yes. Shopify, Cafe24, WooCommerce, Medusa, HubSpot, Salesforce, custom Node / Python / Rails backends. We’ve integrated with all of them. Send us your stack in the first email. We’ll tell you honestly if there’s a constraint that would shape the project.
03 / 05
Michael (CEO) handles strategy and client calls. Nick (AI Generalist) handles design, engineering, and 3D. Network collaborators are introduced by name and stay on the project through launch. No account managers, no senior-staff rotation, no translation layer.
Small core team in Busan and Seattle, with a vetted network of collaborators (3D artists, WebGL engineers, motion designers, Korean copywriters, print specialists) who join projects based on scope. We don’t publish headcount because it changes project-to-project. When a project needs ten people, ten people show up.
It’s Nick’s actual role. The person who can write the code, design the interface, model the 3D asset, integrate AI into the production pipeline, and run operations. All on the same project. Not a specialist; not a manager; a hands-on operator who uses AI as a force-multiplier across traditionally separate disciplines. It’s why our timelines are shorter than a studio of our size “should” be able to deliver.
Busan is UTC+9; Seattle is UTC-8; Bangkok (where Nick is often based) is UTC+7. Between us, we cover Asia and North America comfortably. European clients get most-of-a-day overlap. We run async-first communication (Slack, Notion, email) with weekly synchronous calls.
04 / 05
Every proposal has a single fixed number with a named hour cap. We scope based on the feature list, integration complexity, and target quality, not on a daily rate. You know the number before you agree to anything.
Because most agencies don’t, and it creates friction for everyone. You can scope your project mentally against our published ranges before the first call. And we don’t have to spend the first week of every engagement negotiating price floors.
Productized sprints are fully fixed. Custom engagements are fixed per-proposal (single price, named hour cap) but the price varies with scope. Every proposal has one number, not a range.
Usually no. For most projects we scope inside the proposal from the first 15-minute call. For one in ten projects the discovery is real work, and we scope a paid two-week discovery sprint ($5K to $10K) with a named output: a research deck, a technical spec, or a working prototype.
Standard: 40% at kickoff, 30% at mid-milestone, 30% on launch. For sprints under $10K, 50% / 50%. For projects above $50K we’ll sometimes split into 30 / 30 / 40 or milestone-tied payments. For retainers, monthly in advance.
Rarely. We don’t run promotions. We do occasionally price aggressively on a first project with a client we believe will become a long-term relationship. It’s a judgment call, not a policy.
No. If you need to see our thinking before hiring us, start with an AI Ops Audit ($5K) or a single Web 3D Product Showcase ($6K). Both are real work with a real output.
We operate from Busan, Seattle, and Bangkok. Our overhead is materially lower than an agency running West End offices and a full account-management layer. We pass the difference on as pricing. Quality bar is the same.
05 / 05
We integrate AI tools into specific production steps where they produce measurable time savings without quality loss. 3D baseline geometry (Tripo, Hunyuan3D, Meshy). Copy drafting and localization (Claude, GPT-4, DeepL). Design exploration (Stitch by Google, Midjourney). Code scaffolding and review (Cursor, Claude Code, Copilot). QA automation and testing. The output is what matters, not the tool.
No. The final craft decisions (typography, interaction design, 3D art direction, copy voice, information architecture) are human. AI compresses the production time on well-defined tasks. It does not replace taste.
No. We don’t use client work for training data. Contract language available on request.
Not the final asset. AI accelerates the base mesh, the material exploration, the iteration speed. The finished 3D asset is the product of human direction and CAD or reference data from the client.