Spline AI

What is Spline AI?

Spline AI is the generative AI feature integrated into Spline’s interactive 3D and 2D design platform. Users can create 3D objects from text or images, generate seamless textures, convert 3D scenes into images in various styles, and then continue to adjust the geometry, textures, lighting, animation, and interactions within the Spline editor.

Spline is not only a 3D model generator; it is also suitable for use in websites, applications, product demonstrations, and interactive experiences. The resulting scenes can be embedded in web pages, or integrated into products using React, Next.js, Swift, and Kotlin.

Convert text to 3D text to create 3D models

After the user enters descriptions of the object, shape, material, and style, the AI generates four potential versions, from which one is chosen to be converted into a 3D model. It is suitable for quickly creating conceptual props, cartoon characters, furniture, and decorations.

The grids that are generated may have issues related to topology, scale, and detail, and they may not be suitable for use in games, animation rigging, or manufacturing. For professional applications, it is necessary to further refine the grids, UV mappings, and materials.

Convert image to 3D image to generate 3D.

Image to 3D uses a front-facing 2D image as a reference to convert characters, products, or objects into 3D models. The cleaner the background of the reference image, the more straight the viewing angle, and the more complete the outlines, the more stable the resulting 3D model will usually be.

A single image cannot provide complete information about the back side of an object or about areas that are obscured; therefore, AI will have to estimate those invisible areas. This feature cannot be relied on for an accurate representation of the product or for obtaining its exact dimensional specifications.

Remixes and variants

Users can continue to remix the objects that have already been created, maintaining the general direction while exploring new colors, shapes, and details. The satisfactory results can be saved in a personal 3D library for reuse in various scenarios.

AI Texture generation

AI Textures can create seamless textures based on text, such as brick walls, concrete, water surfaces, and mosaics, which can then be applied to 3D objects. After creation, it is still necessary to check the seams, proportions, normal orientation, and how the textures appear under different lighting conditions.

AI Style Transfer

Style Transfer is an experimental feature: users first create a 3D scene, then choose styles such as cartoon, clay, or fluid, along with elements like rain, snow, night, or sunset to enhance it, and finally refine the output image using text.

It generates stylized 2D images based on 3D structures; this is not the same as permanently converting a complete 3D scene into another set of editable materials. The stability of the experimental functions and the relevant licensing details should be referred to in the product instructions.

Omma natural language construction

Spline also supports the Omma approach, allowing users to create interactive experiences, websites, 3D models, and applications using natural language. It is suitable for quickly developing prototypes, but the AI-generated interactions, responsive layouts, and performance need to be tested manually.

Hana AI Image Workflow

Hana is an infinite 2D canvas introduced by Spline; it allows for the creation of images from text or pictures, and it supports the use of various models such as Flux, Nano Banana, and Seedream. Users can use up to six pictures or frames as references in order to combine, modify, and create variations of these images.

Hana also offers background removal, scaling, and the creation of frame references; it enables vector shapes, text, and layout guidelines to influence the AI’s output, and the generated images can be incorporated into events and effects.

3D editor

After the object is created, the user can continue to edit its dimensions, position, rotation, material, color, lighting, and camera settings. Basic modeling and scene tools can also be used to create a complete experience. AI helps to speed up the initial stages, but the ultimate quality still depends on the design and optimization efforts.

Materials, lighting, and environment

Spline supports multiple material layers, textures, lighting, shadows, and environmental settings. Web projects should manage high-resolution textures, transparent materials, as well as the number of real-time shadows and lights, in order to avoid overloading mobile devices.

Animations, States, and Events

Objects can create states and transitions, and they can respond to events such as mouse clicks, touches, keyboard inputs, scrolling, as well as commands like Start, Look At, and Follow. It is possible to create hover animations, product interactions, and scroll-based narratives without the need for traditional 3D programming.

Variables, APIs, and Webhooks

Pro offers unlimited Variables, APIs, and Webhooks, and it also supports local storage of variables. It enables the updating of text, numbers, and status values based on data from interfaces, making it suitable for real-time dashboards, configurators, and interactive products.

External APIs should pass through a secure backend proxy; keys must not be exposed in public environments. Additionally, cross-domain issues, retry mechanisms for failures, data validation, and rate limiting need to be taken into account.

Difference between Real-time API and Code API

The Real-time API is used to connect external APIs within the Spline editor; the Code API enables developers to read variables from their own web page code, modify object positions or materials, trigger transitions, and monitor Spline events.

They have different uses.

Web embedding

Scenes can be embedded in web pages using code snippets. The Free version exports content with a Spline watermark; Hobby removes the watermark from the exported web content. The Pro version not only removes the Web Embed watermark but also provides more extensive development capabilities.

Before embedding, test the loading of the initial screen, lazy loading, performance on low-power devices, touch interactions, and alternative accessible content. Complex real-time 3D graphics increase the workload on the CPU, GPU, and network.

React and Next.js

The official React components encapsulate the Spline Runtime, enabling the loading of scenes, event listening, and invocation of object methods in React and Next.js applications. The Code API allows for the retrieval of objects, adjustment of scaling, triggering of events, and dynamic modification of properties.

Server-side rendering and route switching require proper handling of dynamic imports, component unloading, and scene caching, in order to avoid errors on the first page or unnecessary consumption of resources.

Export to Apple platform

Pro supports export in Apple’s format; it allows scenes to be integrated into iPhone, iPad, Apple Silicon Macs, and Vision Pro using Swift code snippets or the generated Xcode projects. The official runtime is based on Metal and SwiftUI.

The current native runtime requires iOS or iPadOS 16 or later, along with a chip of model A13 or higher; for Macs, macOS 13 and Apple Silicon are required, while Vision Pro needs an M2 chip.

To publish on the App Store, a developer account is also required.

Export to Android platform

Spline can generate Kotlin integration files as well as APK or AAB files, thereby enabling the inclusion of interactive 3D elements in Android applications. Performance varies significantly depending on the GPU and system version, so testing is necessary on mid-to-low-end devices.

Export of 3D files and videos

The platform supports exporting in formats such as USDZ, STL, GLTF with colors and textures, as well as videos; the specific formats available depend on the package chosen. Format conversion may result in the loss of events, material layers, or special effects, so it is necessary to verify the exported files in the target engine.

Comparison of Spline packages and prices

The paid plans allow for an increase of 1,000 AI Credits for every $5 per month; the maximum number of additional credits that can be acquired varies depending on the plan chosen. The price is calculated based on the number of seats, while taxes and the currency used depend on the payment page details.

How to choose between Hobby and Pro?

Hobby is usually sufficient when you need unlimited files, watermark-free web exports, and additional resources; for watermark-free Embed options, native mobile exports, unlimited variables, as well as API and Webhook capabilities, you should choose the Pro plan or a higher-tier plan.

Max and Enterprise

Max is designed primarily for users who engage in frequent AI-generated content creation, and it includes 10,000 Credits. Enterprise adds features such as SSO, multi-team management, version history, code export, and self-hosted export, making it more suitable for organizations with specific security and deployment requirements.

Performance optimization

For 3D web pages, it is necessary to reduce the number of polygons, material layers, texture sizes, real-time lighting effects, and animations that play simultaneously; lazy loading and on-demand rendering should be employed. Testing should also be carried out on mobile devices, in battery-saving mode, and with integrated graphics cards.

Data and AI Privacy

To generate content with AI, it is necessary to send the prompt words and reference images to a cloud service. Before uploading customer products, unreleased characters, or sensitive materials, one should check Spline’s policies regarding the use and management of AI data.

GitHub and the open-source status

The Spline core browser editor, AI-based 3D generation tools, Hana, as well as account and cloud service features are not open-source products. The official GitHub repository provides React components, packages related to the Web Runtime, the iOS runtime, and example projects.

React-Spline is licensed under the MIT license; however, its use in a public environment does not mean that the full editor or AI models are made available under an open-source license. The enterprise self-hosting option also does not provide access to the source code of the core platform.

Spline AI Usage Guide

Complete a basic task.

1. Define the code tasks to be carried out in Spline AI, the scope of the repository, and the acceptance criteria.
2. Connect to or import the test project, and first back up the current branch;
3. First, use Text to 3D to create a 3D generation plan, and then verify the modified file.
4. Use Image to 3D to generate 3D models and make minor adjustments;
5. Run tests, static checks, and builds; unverified code is not accepted directly.
6. Manually check permissions, keys, dependencies, and handle exceptions before merging;

Create reusable professional workflows

1. Select a low-risk, real-world project as a template;
2. Fix the order of use for Text to 3D for generating 3D models, Image to 3D for creating 3D images, as well as Remix and its variants.
3. Record the environment, model, prompts, and failure conditions;
4. Set up manual approval for write, deploy, and delete actions;
5. Compare speed, cost, test pass rate, and the amount of rework;
6. Expand to a team or production environment only after stability has been verified;

Which users is it suitable for?

• Web designers who create interactive 3D websites;
• Visual creators who can rapidly generate 3D conceptual assets;
• The team responsible for creating product configurators and brand interaction pages;
• Developers who integrate 3D elements into React and Next.js projects;
• A team that creates native 3D experiences for iOS, Android, and Vision Pro;
• Designers who need 2D AI images to explore together with 3D scenes.

Product advantages

• 3D, 2D, animation, and interactivity are handled in the browser;
• It supports the creation of 3D images, as well as textured and stylized images from text and pictures;
• Events and states can be created without traditional 3D programming;
• Covers Web, React, Next.js, Apple, and Android;
• It offers a free version along with a clear tiering of AI Credits;
• The official team provides multiple runtime environments and development components.

Restrictions and Precautions

• The grid precision and topology of AI-generated models are limited, and StyleTransfer remains an experimental feature.
• Complex web scenarios can lead to performance and loading issues, while features such as code, self-hosting, and version history are available only in the Enterprise version.
• Payment is based on the number of seats, and AICredits may also incur additional costs;
• The commercial rights related to the models, reference images, and the generated assets need to be checked separately.

Frequently Asked Questions

Is Spline AI free?

Free allows you to get started and experience limited AI features; for high-frequency generation and watermark-free publishing, Hobby, Pro, or Max is required.

Can Spline AI convert images into 3D?

Yes, Image to 3D uses the front-side image to influence the generation, while the back side and covered areas are inferred by AI.

Can splines be embedded in React?

Yes, the official platform provides React and Next.js components as well as a Code API.

How much is Spline Pro?

The reference price is $20 per seat per month when paid annually, or $25 per seat per month when paid monthly.

Is Spline open source?

The core editor and AI platform are not open source; however, certain components such as the official React, Web, and iOS runtimes are available publicly.