The gear expert you can hire, on every product page.
Your customers ask real questions before they buy — will this fit my rig, how do I save a preset,
what do I need with it — at 11pm, on a Sunday, on a product page that cannot answer them.
GearPlug answers, in your store, from the manufacturer's own manual.
And it's a brilliant shopper. It reads the page it sits on — your copy, your specs, your price —
then shops the store alongside your customer: the conversation follows them from
product to product, every answer links gear from your catalogue — good, better, best, with prices —
it remembers what they already own, and it says so plainly when it doesn't know rather than
inventing an answer.
One script tag. No backend work. Built for music retailers, marketplaces and manufacturers.
Reads your page content · Recommends from your catalogue · Knows the customer's rig · One script tag
1,622
Validated Manuals
And growing
83,409
Manual Pages Indexed
And growing
45+
Studio & Gear Specialists
The Short Version
If you own the P&L rather than the codebase, this is the whole proposition on one screen.
WHAT
A product expert that scales
A chat assistant that sits on your product pages, trained on the manual library and still being extended. It answers the technical questions your customers ask before they buy — the ones that currently reach a human, or go unanswered.
HOW
Your content plus our library
One script tag. It reads the page it is on — your description, your spec table, your price — and answers from that first. Anything your page does not cover, it retrieves from the manual library. Where neither applies, it says so rather than inventing an answer.
WHY
Nobody can staff this
You cannot hire enough people who know a Sub 37, a Scarlett 2i2 and a 1176 well enough to answer at 11pm on a Sunday. The knowledge exists — it is sitting in manuals nobody reads. This is the only way to put it on every page at once.
IMPACT
Four places it shows up
Pre-sale questions answered without a human. Complementary gear attached to orders that would have been single-item. First-party data on what your customers already own. And answers available in every timezone, permanently.
How We Got Here
Twenty years of selling gear, teaching it and using it in studios — and one
problem that never went away.
1
Behind the counter, then behind the desk
I spent my youth DJing, and went on to work behind the counter at Guitar Center, Brook Mays Pro Shop and West L.A. Music. From there I went deep on music software and ended up teaching Logic Pro for Apple. Then years in Los Angeles studios — countless sessions, alongside some genuinely extraordinary artists, a few you would know by name.
2
The same thought, over and over
The knowledge in those rooms never reaches the people buying the gear. The engineer who knows why that compressor sits better on that vocal. The producer who knows which two boxes actually talk to each other. That expertise exists — it is just never in the room when someone is deciding what to buy.
3
A guru in your pocket
I wanted to build something that could teach, guide and shop with you. For years it simply was not possible. So I built a prototype: social, between friends — share your gear, talk about it, and have your own studio guru on hand. You taught it what you owned, it read the manuals properly, and you got straight to the part that mattered. I invited friends in to test it. Then life pivoted, and GearPlug pivoted with it.
4
What if it lived on the store instead?
What if the assistant sat on the product page itself? What if it knew the manuals page by page rather than roughly? What if it could walk you through building a preset? What if it knew what you already own and suggested what would genuinely complement it — at your price point, for the music you actually want to make? And what if it brought you back after the sale, at 1am, when you are stuck?
5
Well — I built it
It reads the page it is sitting on, answers from the manual behind the product, recommends from your catalogue rather than the open web, and says plainly when it does not know. That last part was the hardest to build. It is also the part that makes everything else worth trusting.
Don't take our word for it — go and try to break it
We built a working demo storefront so you can use the assistant the way a customer would.
Open any product, ask it something awkward, and see whether it answers from the manual or
tells you it doesn't know.
Open the Demo Store
A real storefront with real gear and real manuals behind it — the prices and stock are
made up, the answers are not.
The Entire Integration Is One Line
No API build, no data feed, no migration, no engineering quarter. Your team pastes a script tag into the product template and it is live across the catalogue.
PASTE ONCE, INTO YOUR PRODUCT TEMPLATE
<!-- GearPlug AI Widget — same tag on every product page -->
<script
src="https://gearplug.ai/widget/gearplug-widget.js"
data-key="YOUR_KEY"
data-color="#c026d3"
data-catalog-url="/search?q={query}">
</script>
That's the whole integration — the same tag on every page, not one per product. The widget reads which product it is sitting on from the page itself: your OpenGraph tags, JSON-LD, or the page title, the metadata your store already publishes. data-catalog-url is the important one: every product the assistant recommends links into your search results, so a recommendation ends on your site. For the rare page that needs it, per-product overrides exist — data-product names the product explicitly and data-manual pins the exact manual — but a template-driven store never touches them.
WHAT IT DOES ON THE PAGE
Product-Aware Chat
The widget knows which product page it's on, and reads your own description and spec table off the page. Anything your listing already answers comes back instantly; anything it doesn't goes to the manual.
Attaches the Rest of the Order
Customers tell it what they already own. It works out what is missing — the interface, the right cable, the stand, the expansion — and every suggestion links into your catalogue search, not away from it.
Grounded, and Honest When It Isn't
Specs and settings come from the manufacturer's own documentation. If no manual backs an answer, the widget says so instead of inventing a page number to cite.
Knows the Music, Not Just the Specs
Which records were made on it, which artists used it, what it is worth on the used market, what people usually pair it with. The specialists behind it cover music history and production, not only technical documentation.
How It Works for Retailers
An afternoon of work for one front-end developer, then it runs
1
Paste the Script Tag
Into your product template, once. It identifies the product from metadata you already publish, so it works across the whole catalogue without a per-product build. Shopify, Magento, BigCommerce or bespoke — it is a script tag, so it does not care.
2
It Reads Your Page, Then Our Library
Questions your own page already answers come straight back from your content. Everything deeper — a parameter buried on page 40 of a manual, whether two devices sync — is retrieved from the library. The customer never has to know where the answer came from.
3
You Get the Questions and the Rigs
Every conversation is a customer telling you, unprompted, what they own and what they are trying to do. That is merchandising and buying intelligence you currently have no way of collecting at the point of sale.
It Shops With Them, Not Just Answers Them
A chat box answers questions. A shopping companion walks the store with the customer — and this one does.
The Conversation Follows Them
Ask about one synth, click through to its rival, and the thread comes along — the assistant knows which page they're on now, keeps what was said before, and compares the two without being re-told. Browsing with it feels like walking the aisles with the best person on your floor.
Every Answer Is Shoppable
Ask what gear built the Motown sound and the history lesson arrives with the modern equivalents linked into your catalogue — a buyable Precision Bass beside the story of one. Ask a which-should-I-buy question and it lays out good / better / best with street prices. Chat becomes cart.
Questions Worth Asking, Dealt Fresh
One tap deals four new questions tuned to the product on the page — who made records with it, the studio trick nobody knows, what to check before buying used. Shoppers who didn't know what to ask discover the product has more to say.
A Historian Behind the Counter
Which records it was on, which artists swore by it, why it became a classic — the lore that sells an instrument, told at the moment of consideration, with the manual's precision underneath when the question turns technical.
What It Changes in the Business
Four line items. Three of them you are already paying for, in staff time or in carts that were never completed.
The hours your team spends re-answering
Pull a week of pre-sale tickets and live-chat transcripts. A large share will be the same handful of questions: does it work with my interface, what is the difference between these two, does it come with the power supply, can it do this one thing. Every one of those is answerable from a manual or from your own product page, and every one currently costs a person's time.
Orders that should not have been single-item
Someone buying their first synth needs a sustain pedal, an interface, a stand, cables and a case, and does not know it yet. A floor salesperson catches that. A product page does not. The assistant does, because it knows both the product and what the customer told it they already own — and each suggestion links into your own catalogue search.
First-party data on what your customers own
To get a compatibility answer, customers volunteer their actual rig. Not a survey, not an inferred segment — an unprompted list of the gear in the room, attached to a live purchase intent. That is buying, merchandising and bundling intelligence you have no other way of collecting at the point of sale.
The hours you were never going to staff
Most gear research happens at night and at weekends, which is exactly when your expertise is offline. Traffic in other timezones has the same problem permanently. This is the only version of your product expert that is awake for all of it.
We do not publish uplift percentages. Any number we quoted would come from someone else's catalogue, traffic and margin, and would tell you nothing about yours. Run it on one category, leave the next one alone, and compare them — that figure is worth more than anything we could put on this page.
Who This Is Built For
If you sell, list, or make music equipment, the same problem shows up in a different place
Online Retailers
You carry thousands of SKUs and no floor staff who knows all of them. Put an expert on every product page that answers from the actual manual — and points to the cable, interface, or stand that product needs.
Marketplaces
Your sellers write two-line listings for forty-year-old gear. Give buyers something that can explain what they're actually looking at and whether it fits the rig they already own.
Manufacturers
Your documentation is thorough and nobody reads it. An FAQ section answers twelve questions; this answers the long tail, out of the manual you already wrote.
Built With Care, Not Taken
This is a research lab for gear knowledge, and it treats manufacturers' documentation with respect: sourced only from what they publish, robots.txt honored everywhere, provenance recorded for every document, and any manufacturer can have theirs corrected or removed with one email. Every manual is checked against the product it claims to document before it answers a question — wrong-model PDFs, duplicates and image-only scans are held back rather than quietly answering.
Pilot partners' catalogues go to the front of the sourcing queue. Missing a product you sell? Send the list.
What's Inside the Engine
Curated Knowledge Base
1,622 equipment manuals across 212 manufacturers — 83,409 pages, hand-sourced and validated before indexing. Manufacturer documentation, not scraped PDFs of unknown provenance.
It Knows the Music, Not Only the Manual
Mixing and mastering engineers, guitar and drum techs, repair engineers, DAW specialists, genre producers and music historians — 45+ specialists, with intent classification routing each question to the right one. So "what records was this used on?" and "what do people usually pair it with?" get a real answer, which is how gear actually gets sold.
Reads Your Site, Not Just Our Library
It takes your own page content — description, spec table, price — as authoritative, and layers the manual library underneath for everything your page does not say. Your merchandising stays the first answer; the library handles the long tail behind it.
Citations You Can Check
Answers are written from retrieved manual content, then audited: any reference the engine cannot trace back to a section it actually pulled is removed before the customer ever sees it.
The Questions That Stall a Sale
These are the four shapes almost every pre-purchase question takes. The engine routes each one differently.
"CAN IT DO THIS?"
"How do I program a bass patch on the Juno-60?"
Routed to: Vector search across the Juno-60 manual
Answers from: The pages that cover patch programming, quoted as written
Why it matters: The customer stops tabbing between your page and a PDF on a forum
"WILL IT WORK WITH MY SETUP?"
"I run a Scarlett 2i2 and Ableton — will this fit?"
Routed to: Retrieval across both the product's manual and the gear the customer named
Answers from: I/O, sync and connection sections of each
Why it matters: This is the question that ends in an abandoned cart or a return
"WHICH ONE SHOULD I BUY?"
"What's the difference between the TR-808 and TR-909?"
Routed to: Multi-manual retrieval, then a comparison pass
Answers from: The spec and voice-architecture sections of both manuals
Why it matters: A confident comparison closes; a vague one sends them to a competitor
"WHY ISN'T THIS WORKING?"
"My compressor is making the mix sound worse."
Routed to: Troubleshooting intent, matched to the right expert persona
Answers from: General technique, clearly flagged as such when no manual covers it
Why it matters: A post-sale question your support team is answering by hand today
Built to Sit on Someone Else's Storefront
Keys Scoped to Your Domains
Every embed uses a key locked to the origins you name, with per-key rate limits and a daily token budget. A leaked key can't be run up on someone else's site.
It Can't Break Your Page
The widget renders inside a shadow root with styles fully isolated, so it inherits nothing from your CSS and leaks nothing into it. No build step, no framework, no conflicts.
Answers That Check Themselves
A validation pass runs before any answer reaches the customer and strips every page reference that isn't backed by a manual section the engine actually retrieved.
The Difference
Why Not Just Use a Generic LLM?
Generic AI chatbots hallucinate specs, invent features, and give dangerous advice about equipment. The GearPlug engine is different.
GENERIC LLM
- Hallucinated specifications
- No source attribution
- Stale training data
- No user context
- Generic responses
GEARPLUG ENGINE
- Manual-verified specifications
- Source pages cited
- Continuously updated knowledge
- Owned-gear-aware responses
- Expert persona routing
THE RESULT
- Accurate technical answers
- Trustworthy gear advice
- Real purchase confidence
- Reduced support tickets
- Higher engagement
Frequently Asked Questions
How does the widget get added to our site?
One script tag on your product pages. No backend integration required. The widget auto-detects the product from your existing page metadata (OpenGraph, JSON-LD, or data attributes). It works on any website — Shopify, custom builds, WordPress, anything.
How is this different from a regular chatbot?
A general-purpose chatbot answers gear questions from whatever it absorbed during training, and it will confidently invent a spec rather than admit it doesn't know one. This engine retrieves from a library of real equipment manuals before it answers, and a validation pass strips any page reference that isn't backed by a section it actually pulled. When there's no manual to lean on, it tells the customer that rather than guessing.
What does "gear-context aware" mean?
Customers tell the widget what gear they already own, and it factors that rig into every answer afterwards. "Will this work with my interface?" gets a real answer because it has both manuals. It then works out what is missing from that setup and suggests it — and with data-catalog-url configured, each suggestion opens in your own catalogue search.
How many products does it cover?
The library currently holds 1,622 validated equipment manuals — 83,409 pages — across 212 manufacturers, spanning current production gear and long-discontinued classics. Send us your catalog and we'll tell you honestly what's covered today and source what isn't.
Can it be branded to match our site?
Yes. data-color sets the accent to your brand, and data-catalog-url points recommendations at your own search results, so a suggested product opens on your site rather than anywhere else. It renders in an isolated shadow root, so it inherits nothing from your CSS and cannot disturb your layout — it reads as part of your site, not a bolted-on third-party tool.
How fast are the responses?
Fast enough that customers keep typing rather than leaving. Behind each answer the engine classifies the question, retrieves the relevant manual sections, generates a response and validates it before anything renders — and the widget streams a typing indicator throughout, so the page never looks stalled.
What happens when there's no manual for a product?
It answers from general expertise and says that's what it's doing, rather than inventing a page number. A validation pass runs on every response and removes any citation that isn't backed by a manual section the engine actually retrieved — so a cited page is always a real one. If a product matters to you and isn't covered, we'll source and validate that manual.
We sell far more products than you have manuals for. Does that break it?
No, because the library is only half of it. The assistant reads the page it is on first — your description, your spec table, your price — so on a product with no manual in the library it still answers from your own content, and still knows what the customer owns and what complements it. The manual library is what lets it go deeper than your page on the 1,400+ products it does cover. Send us your catalogue and we will tell you exactly which those are, and source the ones that matter to you.
Will it send our customers to a competitor?
Not when data-catalog-url is set. Recommendations then resolve to your own catalogue search, so every suggested product opens on your site. Setting it is part of a standard install, and it is the first thing we check on a retailer deployment.
What does it cost?
Pricing depends on catalog size and query volume, so we quote it per retailer rather than posting tiers. Every key carries a rate limit and a daily budget you agree up front, so there's no surprise bill. Tell us a rough SKU count and we'll come back with a number.
Will it slow down or interfere with our pages?
No. It's a single script with no framework or build step, and it renders inside a shadow root — its styles are fully isolated, so it inherits nothing from your CSS and leaks nothing into it. Nothing on your page moves.
See It on Your Own Gear First
Try the widget on a live product page and ask it something you'd expect it to get wrong. If it holds up, send us a handful of your SKUs and we'll tell you what the library already covers.
A person reads every one of these — usually the one who built it.
GearPlug is a working lab for gear knowledge — a proof of concept that grows
every week. Early partners shape what gets covered next and lock in first-year terms.