Tags: posts polarity-music Bitwig Tutorial VST-Plugin AI

Infinite Crate - Generate Streams of Sound in Your DAW

Tutorial | Sep 17, 2025

This video reviews the Infinite plugin, a free AI-powered tool, likely made by Google, that generates an endless stream of music based on user prompts and context, allowing detailed control over the composition process. The creator acknowledges concerns about AI, such as data usage from artists and environmental impact, but is interested in exploring its creative potential and hopes for future open-source or local versions. Overall, the plugin inspires fresh ideas and background textures for musicians while letting them directly shape the generated output, despite some technical limitations and the broader uncertainties about AI in music creation.

You can watch the Video on Youtube

Short Overview

I've been exploring this new plugin called Infinite-Crate, which generates an endless stream of music using AI based on your prompts and the context of the music you've already created. While I share concerns about big companies using artists' data, environmental impact, and the broader implications of AI, I'm also curious to see what's possible with the latest technology. The plugin is free to use, you just generate an API key and start crafting music instantly, tweaking prompts, influences, and key settings as you go. It's a great tool for getting fresh inspiration, and even though it's not perfect and relies on servers for now, I see a lot of creative possibilities, especially if something like this becomes available for local use in the future.

Introduction to the Infinite-Crate Plugin

In this video, I explored the Infinite plugin, a tool created by Google that generates an endless stream of music using AI. The plugin takes prompts and even previously generated music as input to continuously craft new music in real time. While I have my reservations about the broader implications of AI in music, such as data harvesting from small artists, environmental concerns, and how large companies profit from these technologies, I am nonetheless curious to see what is currently possible with this new technology. I want to find out how it works, what we can do with it, and what it means for musicians and producers today.

Getting Started and Technical Overview

The Infinite-Crate plugin is completely free to use, although it requires generating an API key the first time you run it. Once set up, you simply type a prompt, and the plugin generates music based on your instructions. Prompts can be anything from a plain keyword like "ambient" to more descriptive instructions such as "moon landing" or adding certain instruments.
On the interface, there are sliders that control how much your prompt and the context of previously generated music influence the output. This means you can generate a continuous audio stream that evolves based on your real-time input or a chosen seed, balancing between your original prompt and the AI's ongoing generation.

Using Prompts and Influences

One of the unique features is the prompt and context influence controls. The prompt influence slider determines how much your written ideas shape the generated music. The context slider lets you decide whether the next section of music follows from what has already been created or sticks closely to your prompt. You have granular control over the creative direction here.

You can also set parameters like BPM and key. However, these are not precise. For example, specifying a key merely adds it to the prompt, such as "ambient E flat," rather than accurately generating music strictly within that key. Sometimes it helps to correct pitches afterward using tools like Chroma, which I demonstrate by setting everything to E flat minor and adding effects like reverb for texture.

Workflow Ideas and Creative Use-Cases

I mainly use the Infinite-Crate plugin as a source of inspiration and raw material. By recording its output using something like a rolling sampler, then layering effects or harmonizing within my DAW, I can create unique ambient backgrounds or textural pads for my tracks. Because each generation is different, you never get the same result twice, unlike with fixed sample packs or synthesizer presets.

There’s also flexibility to experiment with the prompt by adding things like "angel voices" or "deep bass." The plugin can create layers such as percussion, vocals, or intricate melodies. Adjusting the density control can influence how busy or sparse the music becomes, giving you further customization power.

Comparisons and Future Potential

Compared to other AI music generators such as Suno, which tends to generate complete songs from prompts, the Infinite-Crate plugin feels more like a musician’s assistant: it lets me steer the vibe and texture in real-time, cut and paste sections I like, and use the results as creative building blocks in my workflow.

A key limitation is that music generation currently requires cloud computation , the AI runs on external servers. Ideally, I would want these models to run locally on my own PC, making them faster, more private, and less dependent on third-party providers. Also, a dream feature would be the ability to send audio directly from my DAW to the plugin so the generation could respond even more directly to my own material.

Thoughts on AI and Music Creation

I am aware of and share many concerns within the musician community about the growing influence of AI. It is powerful, sometimes unsettling in its potential, and raises questions about creativity, ownership, and environmental impact. At the same time, these tools can be inspiring and useful, especially for those seeking fresh ideas and new workflows. For me, nothing replaces fine-tuning my own sounds and hands-on music-making, but AI can be a valuable source of inspiration , like a radio station generating endless new soundscapes based on my direction.

Closing Remarks

Overall, the Infinite-Crate plugin is an exciting glimpse into what is already possible at the intersection of AI and music. It might not be for everyone, and it certainly raises debates, but as a free tool, it is worth exploring for curiosity or creative experimentation. Let me know what you think about AI in music, share your views, and stay tuned for more explorations in my next videos.

Full Video Transcription

This is what im talking about in this video. The text is transcribed by Whisper, so it might not be perfect. If you find any mistakes, please let me know.
You can also click on the timestamps to jump to the right part of the video, which should be helpful.

Click to expand Transcription

[00:00:00] Hey folks, welcome back to dates about this plug in here called the infinite
[00:00:03] crate by, I think it's made by Google and it generates an infinite stream of
[00:00:09] music based on your prompt and your context or you're already generated music.
[00:00:15] So it's generating music on the fly with AI and I get it.
[00:00:21] You probably don't like AI like I don't like it.
[00:00:24] And, um, these, all these big companies, you know, getting all the data
[00:00:29] from small artists, never giving anything back and making big money out of it and
[00:00:34] hurting the environment and all that stuff.
[00:00:36] I know it.
[00:00:37] I get it.
[00:00:38] I'm on the same boat, right?
[00:00:39] But at the same time, I want to try out what's new, what's possible right now with
[00:00:45] these type of technologies.
[00:00:47] I just want to give it a try.
[00:00:49] It's not like I'm saying you need to use this or this is the biggest deal ever or
[00:00:53] something like this.
[00:00:54] I'm just giving this a try and see what's possible, what they do with this kind
[00:00:59] of technology and so on.
[00:01:00] So, um, maybe also in the future, we have like open source models or this
[00:01:06] becomes much more efficient on small little PCs or on local PCs.
[00:01:12] And you can generate this on the fly on your own hardware instead of using a
[00:01:18] server.
[00:01:18] So to make use of this, and by the way, this is completely free.
[00:01:22] You don't need to pay anything here for the plugin itself, but you need to generate
[00:01:27] an API key, uh, which is possible if you start the plugin for the first time,
[00:01:32] you generate an API key, and then you can start to create music on the fly.
[00:01:37] Okay.
[00:01:37] So what you do here in this plugin is you just type in a prompt down here.
[00:01:41] We have already your moon landing and base, clarinet, whatever this means.
[00:01:46] And we have some sliders here, which are not volume sliders.
[00:01:49] These are the site, how much influence this kind of prompt has to the output.
[00:01:55] Also on the right bottom here, we have context and prompts for guidance.
[00:02:00] So you can change, uh, how much the context, which is the music we already
[00:02:05] generated with the model and the prompt, which is what we type in your influences,
[00:02:09] the output.
[00:02:10] So you can say, uh, something right down here on the middle, right?
[00:02:13] So the already generated music influence at what comes next, uh, or is generated
[00:02:18] next in the stream, um, and how much the prompt influence at what's generated
[00:02:23] next in the stream.
[00:02:24] Uh, so enough talk, uh, all you have to do is type in something here.
[00:02:28] Uh, we have already something random in there.
[00:02:30] We can also hit random, rich orchestration, snappy snare, whatever this means.
[00:02:35] But I want to generate your ambient add to it.
[00:02:39] And on the left side, you can also decide BPM and the key and the description
[00:02:46] or the manual says, this is actually not very precise, instructs the model
[00:02:50] to generate audio in the specific, uh, specified the key.
[00:02:54] It's imprecise and does not differentiate between major and relative
[00:02:59] minor might need to reset context.
[00:03:02] So I guess when you enable this here and say E flat, uh, it tries to add
[00:03:08] this to the prompt.
[00:03:09] So instead of sending ambient here to the server, we send probably
[00:03:12] ambient E flat to the server.
[00:03:15] So it's just changing the text prompt in the background.
[00:03:17] I, that's my guess.
[00:03:20] So we have ambient here.
[00:03:21] Maybe we want to also add some voices, uh, maybe angel voices add to that, right?
[00:03:32] And then we just hit play here.
[00:03:35] I'd play in the background and then hit play.
[00:03:37] And then it generates stuff that gives you already a nice vibe.
[00:03:50] Maybe acoustic.
[00:03:59] So, and because we have your context and prompt 50, 50 kind of, it generates
[00:04:13] the guitar based on what we already generated.
[00:04:17] Maybe in, make the influence a bit bigger here.
[00:04:22] So it gives you already something to start from an inspiration.
[00:04:33] So what I did then is I used here just chroma, which is this kind of pitch
[00:04:38] correction tool.
[00:04:40] And I say I want to have E flat minor.
[00:04:45] And then just put the reverb on that.
[00:04:48] So I'm using this basically for, for ambient.
[00:04:52] So it was unavailable.
[00:05:09] So there is some, uh, some down times here, I guess a lot of people trying
[00:05:13] this out at the moment, so the server is probably really busy.
[00:05:17] And people already creating 50 tracks per minute.
[00:05:20] Uh, I don't know.
[00:05:22] But, uh, before I tried this out here, it was pretty stable for a long time.
[00:05:27] So we have to hit reset here, connecting to the server, hit play.
[00:05:31] And it starts again.
[00:05:34] So you probably want to record this here, maybe with a, what's the name,
[00:05:39] rolling sampler.
[00:05:42] Yeah.
[00:05:43] Maybe you can also add here.
[00:05:50] Piano tech.
[00:05:53] And then we just clone here, maybe the super massive.
[00:06:00] And here we use something like this.
[00:06:04] Let's hear the sound.
[00:06:09] It's actually in the key because we are using, we are using chroma.
[00:06:12] So instead of using omni-sphere or whatever synthesizer and random presets
[00:06:29] to actually have a base for your ambient track, um, you can just pull up this here,
[00:06:36] generate some random noise sound, put it into the sampler, um,
[00:06:42] or just record it and leave it in the background as a loop.
[00:06:48] And you have a nice background.
[00:06:49] And every time it's something different, you don't need to use, you know,
[00:06:52] presets or new VST instruments.
[00:06:54] You just use this here.
[00:06:56] I don't know if this works.
[00:07:00] Singing voice woman, let's try it out.
[00:07:06] Um, there's also stuff here at the right side, density,
[00:07:19] controls the density of the audio.
[00:07:21] Higher values generate busier sounds.
[00:07:23] Turn this off to let the model decide so we can narrow down here.
[00:07:28] What the model has to generate.
[00:07:30] I guess also this here just influences what you send text-wise to the prompt
[00:07:36] or how you change text prompt to generate something different.
[00:07:39] Let's add some percussion here.
[00:07:58] So my opinion, this is much better as using a Suno and typing in the prompt.
[00:08:16] And then you get a ready-made track here.
[00:08:18] You can decide and make some, you know, you can make some decisions
[00:08:21] and then just cut and paste and make something based off this, add something to it.
[00:08:27] Maybe they add something here like where you can add music or record music
[00:08:34] from your door that you create and then play something on top.
[00:08:39] But this plugin, yeah, this would be nice because you have already
[00:08:42] out of context for the already generated music.
[00:08:45] So it would be nice if we, if I can just stream my music into this plugin
[00:08:49] and then generating something based off of that.
[00:08:52] This would be very interesting.
[00:08:57] Also here in the background, you can record with the rolling sampler forever.
[00:09:02] And you have so much material and you can use this then in the sampler,
[00:09:07] make pad sounds out of it, I don't know, lead sounds or whatever,
[00:09:12] instead of just using sample libraries and omnisphere presets or whatever.
[00:09:18] So my opinion, really dope.
[00:09:20] And I really hope we get something like this with the local model
[00:09:26] where we can generate this locally on the local PC instead of using servers.
[00:09:31] . Yeah, deep bass.
[00:10:01] Yeah, at the same time, it's it's very scary what you can do with AI nowadays.
[00:10:18] And it's scary where all this is heading and how many options you have.
[00:10:23] And actually, you don't need to do that much, probably any more in the future.
[00:10:26] You just call up such a plug in.
[00:10:28] And then, I don't know, it's probably not fun for a long time.
[00:10:34] I really enjoy making my own sounds, tweaking and twisting knobs.
[00:10:38] And you know me, you know how I use the grid.
[00:10:41] It's very fun to me.
[00:10:43] But sometimes it's also nice to get some random inspirations,
[00:10:46] have a nice little radio station that gives you a stream of
[00:10:50] random infinite sounds and you can influence here exactly what you want.
[00:10:56] So it's also nice to have that on top.
[00:10:58] I guess I probably end this video here.
[00:11:02] Let me know what you think in the comments down below.
[00:11:04] I know a lot of people don't like AI.
[00:11:06] Let me know what you think in the comments.
[00:11:10] Leave a like, leave a subscription and see you in the next video.
[00:11:14] [MUSIC]
[00:11:24] [MUSIC]
[00:11:34] [MUSIC]
[00:11:44] [MUSIC]
[00:11:54] [MUSIC]
[00:12:04] [MUSIC]
[00:12:12] [MUSIC]
[00:12:22] [MUSIC]
[00:12:32] [MUSIC]
[00:12:42] [MUSIC]
[00:12:52] [MUSIC]
[00:13:02] [MUSIC]