# Delino React Forge v0.2.0

A friend of mine is making a game, so I started trying a few asset creation tasks with React Forge. The game sound effects turned out well, and I was also happy with how the 3D models looked.

Delino React Forge 0.2.0 expands the asset formats you can create with React. In this post, I want to show two of the additions: **procedural sound effects exported as WAV** and **static 3D models exported as GLB or FBX**.

![](https://cdn.hashnode.com/uploads/covers/64c9efa7b925449e35ee2318/9f7de687-7e9e-445c-973b-3238f2fbeac8.jpg align="center")

## Game sound effects

The sound examples below are a single gunshot and a burst of automatic fire. Each plays twice, with no background music or narration, so you can hear the effects themselves.

%[https://www.youtube.com/watch?v=mxDJTkN1HFc] 

React Forge creates these effects through procedural synthesis. Its `Sound`, `Noise`, and `Tone` components let you describe the layers in React and export the result as a WAV file.

The gunshot example combines a muzzle crack, a low-frequency body, a falling tone, a mechanical click, and a decaying tail. Those layers are composed in code, with controls for timing, envelopes, filters, gain, and pan.

The implementation and example code are in the [SFX pull request](https://github.com/delinoio/oss/pull/987).

## 3D models in GLB and FBX

For the visual examples, I asked for high-quality assets that I could inspect visually. The resulting AURA concept set includes headphones, a desktop DAC, a headphone stand, and a combined studio scene.

![](https://cdn.hashnode.com/uploads/covers/64c9efa7b925449e35ee2318/68f0afdb-50e5-4b94-ac8b-c898e98316e1.png align="center")

I was pleased with the detail in these models. The headphones include leather folds and stitching, while the DAC has engraved details, textured controls, and ventilation openings.

%[https://www.youtube.com/watch?v=H1ajZnwb8CY] 

The video shows Blender renders of the exported GLB files, followed by a comparison of the same DAC exported as GLB and FBX. The source models and materials are preserved; presentation cameras, lights, and a neutral floor were added for the demo.

These are static models. The movement in the video comes from the camera, and the current export support does not include rigging or animation.

![](https://cdn.hashnode.com/uploads/covers/64c9efa7b925449e35ee2318/cf983b29-fa25-4fae-b6f9-991c8010a0ba.jpg align="center")

The [3D pull request](https://github.com/delinoio/oss/pull/988) contains the implementation and example code, along with validation of the exported files.

## Visual verification helps

One reason this workflow works well is that today's AI agents can inspect visual output as part of their work. For 3D assets, that lets them check the appearance of the exported model as part of the task.

That was also why I asked for high-quality examples: I wanted results I could look at and judge for myself.

## Use it through MCP

React Forge includes an MCP server in its CLI, so you can connect an MCP-capable agent and create assets through it.

With Node.js 24, install the package in your project:

```sh
npm install @delino/react-forge@0.2.0
```

The local MCP server starts with:

```sh
npx react-forge mcp --cwd "$PWD"
```

Configure your MCP client to launch the server from your asset project. The agent can execute React tasks, inspect and update a session, and explicitly export the resulting files. The [MCP guide](https://oss.delino.io/react-forge/mcp) covers client configuration and the available tools.

You can download the generated sound effects and 3D files from the [public asset folder](https://drive.google.com/drive/folders/1zL6CdHXWnYgYuM-Sl9MOes2aR-gZ3HSQ), and find the rest of the guides in the [React Forge documentation](https://oss.delino.io/react-forge/).