Palatial's biggest growth asset is already public: 1,438 asset pages that Google is invited to index, each with mass, friction, inertia and Isaac Sim validation. But what search and AI engines read is often the raw generation input: a description of "Imported USD asset.", or "bell pepper, realistic, photorealistic textures, accurate scale, single item, accurate real-world scale, no background", and in places internal batch notes. An engineer searching "simready pry bar usd" or "articulated drawer mujoco asset" gets nothing that says what the object is, what it weighs or which simulators it works in. For whoever leads growth, this is the fastest lever: the pages exist, only the words are missing.
First action
Rewrite the titles and descriptions of the 20 most-viewed assets by hand.
How to measure it
Sign-ups starting on an asset page
- Now
- –
- Target
- +100%
- By
- 90 days
Steps
Write the template
Title and description built from name, material, mass, size and export targets.
Growth lead 1 day
Hide pipeline text
Keep prompts, import notes and batch notes out of public fields.
Founding engineer 2 days
Merge duplicates
One canonical page per object; redirect the rest.
Founding engineer 2 days
Add structured data
Name, image, dimensions, mass, file formats, breadcrumb.
Founding engineer 1 day
Add the next step
"Generate a variant" and "Open in Isaac Sim" on every page.
Founding engineer 2 days
Measure
Search Console impressions and sign-ups by landing page.
Growth lead weekly
Why it works
Generate each page's title and description from the data already on it: "<Object>, <material>, <mass> g, <dimensions>: SimReady USD for Isaac Sim, MuJoCo and Newton". Keep prompts and pipeline notes in the dashboard only. Add Product-style structured data, a breadcrumb to its collection, links to similar assets, and one clear action: "Generate a variant of this" with free tokens on sign-up.
Spec searches are long-tail and high-intent, and today the competition there is thin. Every page that answers "is this the object, at the right scale, in my simulator?" becomes a sign-up path into the generator.
Ready to use
CopyAsset page title and description
Title: {name}, {material} | SimReady USD for Isaac Sim, MuJoCo, Newton | Palatial
Description: {name} simulation asset: {mass_g} g, {w}x{d}x{h} mm, {body_type}, collision and physics validated in Isaac Sim. Generate a variant or add it to your workspace.
Never show: generation prompts, import notes, batch or resubmission labels.Risks and how to handle them
Generated text is generic
Use measured fields only; no adjectives.
Duplicate pages split rankings
Canonicalise before resubmitting the sitemap.
Evidence
- Asset pagesEach asset has a public, indexable page on dashboard.palatial.cloud/viewer/, and that sitemap lists 1,438 of them. Many titles and descriptions are the raw generation input: the Cupcake Pan page's description is "Imported USD asset."; library cards read like "bell pepper, realistic, photorealistic textures, accurate scale, single item, accurate real-world scale, no background"; some show internal batch and resubmission notes, and a few assets appear twice under the same name.Source
- Asset pagesA viewer page shows format (USD), body type, mass, density, friction, restitution, centre of mass, inertia and validation ("OpenUSD composition passed", simulator Isaac Sim), with "Public preview; account required for downloads" and "Add to workspace". On the Stanley 7" Wonder Bar II page the description says the object was weighed at 110.0 g ("use as the total mass") while the physics summary lists a mass of 0.13497 kg.Source
- Asset librarylibrary.palatial.cloud offers "1438 assets", with collections "NVIDIA Robotics 1K Dataset" (935 assets, "battle-tested by NVIDIA research scientists in frontier dexterous manipulation research"), "NVIDIA SimReady" (308 assets "augmented with Palatial physics") and "Mad Max" (16). The library's own sitemap lists only 5 URLs: the home, /simready and three collections.Source
- PressNVIDIA opened SimReady Central, a free library of simulation-ready assets on Hugging Face (October 2026); Lightwheel launched SimReadyGen, an agentic generation engine on OpenUSD; Rigyd claims SimReady OpenUSD plus MuJoCo MJCF in about five minutes; DreamVu released a grocery SimReady library from real-world capture.Source