Skip to main content
49m ago

Endless AI Pedal: Redefining Guitar Sound

While I don't harbor animosity towards Polyend's Endless, it won't be finding a permanent spot on my pedalboard. The demand for an AI guitar pedal wa

9 min read7 views5 tags
Originally reported bytheverge

While I don't harbor animosity towards Polyend's Endless, it won't be finding a permanent spot on my pedalboard.

The demand for an AI guitar pedal wasn't necessarily apparent, yet its emergence was inevitable. Polyend, a respected manufacturer known for its niche and distinctive music gear, has taken a significant step in this direction. With a history of crafting grooveboxes based on old-school trackers and multi-effect pedals featuring step sequencing, there was a reasonable expectation that if anyone could successfully implement an AI effect pedal, it would be Polyend.

The Polyend Endless, priced at $299, is a programmable guitar pedal powered by an ARM processor. It integrates with Playground, a suite of interconnected AI agents designed to transform any text prompt into a functional guitar effect. This innovation means that if you envision a unique effect, you no longer need to hope it already exists; you can simply describe it. For instance, a specific combination of effects, like a ring modulator and auto-wah, might lack market demand for a dedicated product, but the Endless allows its creation. While I'm not entirely convinced this precisely aligns with guitarists' deepest desires, it represents a commendable initial effort to merge effect pedal technology with a large language model (LLM).

Among its notable advantages are dozens of free effects available, a user-friendly AI Playground interface, a reasonable price point, and a genuine commitment to "ethical" AI practices.

However, the process of iterating and testing effects can be time-consuming, the firmware exhibits some quirks, and other custom effects pedals generally offer more granular control.

It's important to clarify that the AI processing doesn't occur directly within the pedal. Instead, Polyend has trained a custom LLM to generate effect code, which users can then load onto the device. While advanced users can develop effects themselves in C++, most will opt to download existing free "Plates" (Polyend's term for effects) from the community site or generate them via Playground prompts. Additionally, a physical faceplate can be purchased for $20 to accompany a downloaded effect.

Currently, the Plates gallery features approximately 60 effects, primarily developed by Polyend. This collection spans a wide range, from straightforward saturators to intricate tape loop simulators and guitar synths, even including self-playing drum machines. Personal favorites include "Grunt," a lo-fi octave-down effect; the expansive "Infinite Hall" reverb; and "Tessera," a granular pitch-shifting reverb. "Stardust" also stands out as an immensely rich granular delay, reverb, and tremolo combination that would be exceptionally difficult to replicate elsewhere.

Polyend is also expanding the gallery to welcome third-party contributions, allowing users to create effects in Playground and submit them for inclusion.

Playground is undoubtedly the primary draw of the Endless. It serves as a web-based interface for multiple AI agents working in concert, all trained on Polyend's proprietary effects library. These AI components collaborate to interpret user prompts, select appropriate effect algorithms, generate code from these building blocks, and then validate the code to ensure proper operation without any detrimental sonic artifacts.

For anyone familiar with chatbot interfaces, the Playground web application will feel intuitive. Users describe their desired effect and its controls – which include three knobs, along with short and long presses of the footswitch – and the system responds with a couple of options to materialize the concept.

Typically, three options are presented. Users can simply select one and proceed, or they can refine ideas at this stage, crucially before any code is generated and, more importantly, before incurring any token costs.

Generating effects consumes tokens. The pedal comes with an initial allocation of 2,000 tokens, with additional bundles available for $20 per 2,000 tokens. This initial amount should suffice for creating a few effects. The system's exact mechanics are somewhat opaque, but more complex effects and increased iterations to achieve the desired result will consume more tokens. A simple fuzz might cost only 20 tokens, whereas a granular looper with rhythmically synced glitches could demand around 500.

It's the iterative process, rather than the initial generation, that is most likely to deplete your token balance, especially when aiming for finely balanced or eccentric sonic qualities. Attempting to guide Polyend's Playground to understand precise resonance levels for a bandpass-filtered delay or the exact amount of ring modulation for a particular fuzz can prove frustrating.

For the purpose of this review, Polyend provided 10,000 tokens (valued at $100) for Playground evaluation. In total, I used slightly over 3,500 tokens, yielding three enjoyable effects and several unsatisfactory ones. Had I been purchasing tokens myself, my willingness to continue iterating would likely have been lower. Furthermore, Playground can be slow to generate code, with complexity determining generation times that range from five to over ten minutes per attempt. I often found myself losing patience and abandoning an idea after five or six tries if the desired outcome remained elusive.

Prompt: A clean multi-tap digital delay where the delay lines have narrow highly resonant bandpass filters with subtle modulation.

The effect that presented the most significant challenge was dubbed "Resonant Taps" by the chatbot. My prompt was for "a clean multi-tap digital delay where the delay lines have narrow, highly resonant bandpass filters with subtle modulation." The initial version lacked any control over delay feedback, which, when combined with resonant filters, consistently pushed the effect to the brink of unpleasant self-oscillation. Even replacing the resonance control with a feedback knob didn't resolve the issue.

Prompt: Gently modulating resonant pings that follow the pitch of my guitar, almost like a string section.

After six iterations, I shifted my approach, prompting with a description of the desired sound rather than a specific effect combination. I aimed for "gently modulating resonant pings that follow the pitch of my guitar, almost like a string section." The result, however, resembled a '70s synth experiencing a breakdown – entertaining, but not what I intended. Having spent approximately $7 worth of tokens on this pursuit, I set it aside to explore a more immediately achievable concept, such as a fuzzy, broken CD-skipping effect.

Even this seemingly simpler effect required considerable effort to refine. My initial prompt for "random stuttering and glitching" with "tap tempo" and a blend knob for a ring modulator proved underwhelming. The stutters, even at their slowest, were machine-gun-like and never felt genuinely rhythmic.

It took six generations and specific requests for the stutters to be synchronized to 16th, 8th, or quarter notes before I achieved the desired result. Despite repeatedly asking to increase the ring mod level in subsequent generations, it still feels a touch too subtle. Nevertheless, I am quite fond of the final outcome; it's delightfully chaotic and possesses just the right amount of grit.

Prompt: A sputtering fuzz with light ring modulation and random stuttering / glitching. One knob controlling the amount of fuzz. One knob blending in the ring mod signal, and the other knob controlling the frequency and duration of random stutters. Tap tempo on the left foot switch to make sure the glitches / repeats stay in sync.

There are also several operational quirks worth noting. The Endless can only load one effect at a time. The loading process itself is straightforward: connect it via USB to your computer, where it appears as an external drive, and drag the effect files over. However, despite the pedal automatically rebooting after an effect is loaded, most effects don't function correctly until the device is manually power-cycled by unplugging and re-plugging it. Similarly, it's often necessary to manually eject the Endless and then reconnect it to your computer before loading a new effect. These steps significantly hinder the efficiency of iterating and testing new effects.

Naturally, the sounds I sought to create are complex, involving numerous variables. Yet, this complexity is precisely the point of the Endless. Countless digital delay and fuzz pedals exist that can perform standard functions with greater precision and ease than the Endless. For those simply seeking conventional effects, investing in a small collection of even inexpensive dedicated pedals would be a far more practical choice than relying on AI. The singular compelling reason to acquire the Endless is to attempt to realize a dream effect that has not yet been invented.

However, it's crucial to manage expectations. The effects generated by Polyend's Playground, based on user prompts, are unlikely to rival the meticulously crafted programming of professional audio developers or even dedicated hobbyists. The Endless truly excels as a rapid prototyping tool for ideas that one might further develop, either by directly programming in C++ for the Endless or by utilizing other platforms such as Max MSP or Pure Data. It is regrettable, though, that Polyend does not provide access to the raw code of effects created within Playground for deeper customization.

For exploring complex sonic ideas, alternative modular guitar pedals like the Poly Effects Beebo, the Empress Effects ZOIA, or even the Eventide H90 might be more suitable. These devices allow users to assemble various building blocks to construct elaborate effects patches. While these alternatives come at a significantly higher cost—$449, $549, and $899 respectively—they also boast vibrant communities, stable firmware, and consistently produce top-tier effects. Personally, I am more inclined to dedicate an hour to manually building an effect on my ZOIA than to contend with the intricacies of an LLM.

Finally, the AI aspect itself may, understandably, deter some users. Yet, to its credit, Polyend appears to be making a genuine effort to implement AI as ethically as possible. Beyond developing code solely from in-house creations or open-source components, the company is also striving to mitigate its environmental footprint. While some elements of Playground rely on external AI agents via API, which is largely unavoidable, founder Piotr Raczyński states that the Playground servers are located on-site, not in a third-party data center. Polyend further asserts that these servers are "almost 100% energy self-sufficient thanks to solar power and heat pumps," effectively positioning them as the "free-range, pasture-raised" equivalent of LLMs.

For staunch anti-AI advocates, these efforts will likely do little to alter their perception of the Endless, which may explain why Polyend initially refrained from using the term "AI" during its announcement. However, for those who are more AI-agnostic or have fully embraced the world of LLMs, these ethical considerations might alleviate any potential reservations. The primary strength of the Polyend Endless lies in its affordability and accessibility as a means to create custom effects. Should a robust community form around it, akin to the one supporting the ZOIA, it could become an excellent source of experimental, unique effects. Nevertheless, given the music world's general skepticism toward AI, the emergence of such a community remains a significant "if."

#AI News#Polyend Endless#AI guitar pedal#Guitar effects#LLM
ES
Editorial StaffEditor

The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.

View all posts
Reader feedback

What did you think of this story?

User Comments

Filter:
No comments yet. Be the first to comment!
Continue reading
View all news