AI GUIDEPartner content

Astra and 3D modeling: How AI created an editable steam locomotive model in Blender

In a demonstration, Tom Kercha gave Astra a blueprint of an old steam locomotive and asked it to build a model in Blender. Within a few minutes, the system created a model from 32 295 separate elements, including the boiler, wheels, connecting rods, and rivets. Astra analyzed the image, wrote code, and checked the assembled scene.

Affiliate link: your price stays the same and the project earns a commission.

In a demonstration, developer Tom Kercha gave Astra a blueprint of an old steam locomotive and asked it to build a model in Blender. Within a few minutes, the system created a model from 32 295 separate elements, including the boiler, wheels, connecting rods, and rivets. Astra analyzed the image, wrote code, and checked the assembled scene.

The result was an editable scene, not a finished image.

What Astra was asked to create from the blueprint

Tom Kercha asked Astra to build an old steam locomotive in Blender from a blueprint. The model included a boiler, wheels, connecting rods, and rivets. The scene contained 32 295 separate elements in total.

In this demonstration, Astra converted the blueprint image into a set of objects and wrote code to position them. The number of elements indicates the level of detail in this particular scene, but not the model’s accuracy: it does not show how closely the objects match the blueprint, how the topology is structured, or whether the model is suitable for manufacturing. The demonstration account does not describe these checks.

How Astra created the scene with code

In the process shown, Astra worked with Blender through code rather than controlling the program with simple mouse movements. The system analyzed the blueprint, created and positioned parts, checked the scene, and fixed errors it detected.

The account does not specify which checks Astra performed or what criteria it used to identify errors.

Why editability matters

All the locomotive’s parts remained editable in Blender. This makes it possible to work with individual scene elements rather than just the final image.

The account does not describe how the objects were organized or named, or whether components could be moved to another project while preserving their structure. The total number of elements does not reveal these characteristics. The demonstration confirms that the parts were editable in Blender, but not that the model was suitable for engineering or manufacturing.

How the system checked and corrected the result

The process did not finish on the first try: in the demonstration, Astra checked the scene and fixed errors it detected.

The account of the locomotive demonstration does not specify which automatic criteria the system used. Nor does it say that Astra compared dimensions with the blueprint, checked the strength of parts, or determined whether the design was technically suitable.

In a separate demonstration of a home interior, Astra checked the render and fixed errors it noticed. Examples included the placement of plants, a throw blanket on a sofa, and a metal sink. This episode concerns an interior, not a steam locomotive, and does not show how the system performs in other projects.

Moving the locomotive to a browser and adding animation

The project was later moved to a browser, where wheel rotation and animation were added. The account confirms these changes but does not provide implementation details, performance data, or information about compatibility with different devices.

What the demonstration shows about AI capabilities in 3D

In this project, Astra created numerous objects using Blender, checked the scene, and fixed some errors. The system worked with an editable scene and used code to create and position parts, rather than controlling the program only through mouse movements.

A single demonstration does not prove that Astra can replace an 3D modeling specialist, accurately reconstruct geometry from any blueprint, or prepare models for industrial manufacturing. It shows a specific process, but does not establish that the method is universally reliable.

The account describes how AI created the scene and carried out operations at the user’s request. It does not specify exactly what checks a person performed before evaluating the model.

Who might find this approach useful

The demonstration may interest Blender users and developers of interactive 3D projects: Astra created a scene containing many editable parts. The account describes one project and does not establish that this approach will suit other tasks.

The accuracy of the locomotive model and its compliance with technical requirements have not been confirmed. The description does not provide acceptance criteria.

Frequently asked questions about creating 3D models from blueprints

Can AI build an editable model in Blender?

In this demonstration, Astra created a locomotive model from separate elements that remained editable. The account does not describe how convenient the model’s structure was to work with.

How did Astra interact with Blender?

The system wrote code to create and position objects, checked the scene, and fixed errors. Its work was not limited to controlling the program through mouse movements.

Does the model guarantee accuracy to the blueprint?

No. The demonstration shows that the scene was built from a blueprint, but does not confirm engineering accuracy, compliance with specified tolerances, or that the parts are suitable for manufacturing.

Can this kind of scene be used in a browser?

The locomotive project was moved to a browser, where wheel rotation and animation were added. The account does not provide performance or compatibility data.

Compare models before you start

The service sets its plans, limits and model catalog. If they differ from this article, contact us so we can update it and record a new review date.

Browse models

Affiliate link: your price stays the same and the project earns a commission.

AI for 3D modeling Astra Blender 3D model from a blueprint editable 3D scene Blender automation

SEO Mind42 editorial team

We explore SEO and neural networks in practice: test services on our own projects, verify prices and limits against primary sources, and share things you can put to use the same day.

📚 Reference guide to SEO and AI 🔄 Materials are updated 🕐 Updated: 6 October 2026

Related reading

All in this section →