Part 2 of the free Audio Plugin Coder (APC) walkthrough shows how to use AI for VST plugin development. In this stage, a simple audio effect idea is transformed into a structured plugin brief that an AI coding agent can use to build a real audio plugin.

For this example, we are developing a gritty Vinyl Noise VST plugin based on real turntable recordings. The APC Dream phase combines creative direction, reference audio analysis, parameter design, and AI-assisted ideation to turn the initial concept into a detailed specification for the development phase.

Part 2 — DREAM
/apc-dream VinylNoise · AI-assisted ideation, reference audio analysis, fingerprinting, and parameter specification.

Missed the setup? Start with Walkthrough Part 1.

From an Audio Effect Idea to an AI-Ready Plugin Brief

Creating a professional VST plugin starts long before writing DSP code. A good plugin needs a clear sonic identity, a defined set of controls, reference material, and an understanding of how the effect should behave.

This is where AI-assisted audio plugin development can be useful. Instead of immediately asking an AI coding agent to write a plugin, Audio Plugin Coder first turns the creative idea into a structured development brief.

The result is a much more useful starting point for the later coding stages: the AI has a clear description of the sound, the intended workflow, the parameters, and the technical direction.

The Vinyl Noise VST Plugin Concept

The example for this walkthrough is a Vinyl Noise effect. Rather than creating another generic vinyl crackle plugin, the goal is a gritty, high-fidelity 1960s broadcast character — dusty, imperfect, and based on real turntable recordings.

• Emulate the noise floor, rumble, crackle, and imperfections of real vinyl recordings.
• Use custom turntable recordings as the sonic reference.
• Add optional tape and tube coloration without simply copying familiar stock plugin interfaces.
• Provide an effect amount of roughly 0–100, with the possibility of extending the range beyond 100.
• Keep the interface focused with approximately 10 useful controls instead of a large collection of unnecessary parameters.

Hundreds of loops recorded from an old turntable with a slightly imperfect cartridge were available in the APC workspace. These WAV files became the reference material for the AI-assisted analysis and creative brief.

What the APC Dream Phase Does

The Dream phase of Audio Plugin Coder converts natural-language ideas into persistent project files. This is important for AI-assisted VST development because a future coding session can read the project files instead of relying on a previous conversation or compressed chat history.

The Dream phase can produce files such as:

• creative-brief.md — the plugin concept, sonic character, intended use cases, and creative direction.
• parameter-spec.md — controls, ranges, defaults, and parameter behavior.
• Reference-audio analysis and fingerprinting notes.
• status.json — the current project phase and information required to resume development.

This approach gives the AI coding agent a persistent specification that can be refined before the actual plugin implementation begins.

Using AI to Analyze Reference Audio

Because the goal was to reproduce the character of the actual recordings rather than rely on a generic idea of “vinyl sound,” reference-audio analysis became an important part of the process.

The AI agent created analysis scripts early in the Dream phase and scanned dozens of WAV loops recorded at 44.1 kHz / 16-bit. The analysis identified several characteristics that could influence the eventual VST plugin design:

• Dark and dusty spectral characteristics with relatively hot recording levels.
• A prominent noise floor combined with rumble and surface crackle.
• A nearly mono program signal with out-of-phase high-frequency surface noise.
• Wow and flutter characteristics and distortion-related measurements.
• Large variations in crackle density between individual recordings.

This is one of the advantages of using AI as part of an audio plugin development workflow: the analysis can help turn subjective descriptions such as “dusty,” “old,” or “gritty” into concrete observations that can influence the design of the effect.

Turning Audio Analysis into Plugin Parameters

The reference analysis also led to an additional creative idea: a groove-position control.

Record quality can change as the stylus moves toward the center of a record. As the linear velocity decreases, high-frequency information becomes more difficult to reproduce accurately, contributing to the familiar character of inner-groove distortion.

Rather than treating this observation as background information, APC incorporated it into the creative brief as a potential macro control for the later development stages.

This demonstrates an important principle of AI VST plugin development: the AI is not simply being asked to generate code. Reference material and creative constraints can be used to help define what the plugin should actually do before implementation begins.

How to Use Audio Plugin Coder’s Dream Phase

1. Complete the Part 1 setup.
2. Place your reference recordings and other useful material in the appropriate APC workspace folder.
3. Start the Dream phase with your plugin name, for example /apc-dream VinylNoise.
4. Give the AI agent a detailed description of the sound, intended use, creative direction, and controls.
5. Review the generated creative brief and parameter specification.
6. Answer clarification questions and refine the concept before moving to the planning stage.

/apc-dream VinylNoise

Tip: Be specific about what you do not want as well as what you want. For example, rejecting generic crackle samples, predictable filter stacks, or familiar plugin interfaces can give the AI much stronger creative constraints.

What You Get from the Dream Phase

At the end of this stage, the project contains a much clearer specification for the next phase of AI-assisted VST plugin development.

• A plugin project folder containing the ideation and specification documents.
• A creative brief describing the intended sound and user experience.
• Reference-audio fingerprinting and analysis artifacts.
• A parameter specification defining the controls and their intended behavior.
• Project status information pointing toward the next development phase.

Why This Matters for AI VST Plugin Development

One of the biggest challenges when using AI to create an audio plugin is giving the coding model enough context to produce something useful. A short prompt such as “make a vinyl plugin” leaves too many decisions open.

A structured workflow provides the AI with a much better foundation. The sonic references, creative brief, parameter specification, and technical observations become persistent project information that can be used by later AI coding sessions.

This makes Audio Plugin Coder more than a simple code-generation prompt. It provides a workflow for taking an idea from concept and reference material toward an actual VST audio plugin.

Next: Planning the VST Plugin

Part 3 moves into the /apc-plan phase. The next stage looks at the plugin architecture, including WebView versus Visage, DSP architecture, implementation decisions, and potential high-risk areas before coding begins.

If you want to follow the complete process of creating a VST plugin with AI, continue with the next part of the Audio Plugin Coder walkthrough when it becomes available.

Get Audio Plugin Coder:
Audio Plugin Coder on GitHub

Series:
Part 1 – Setup