web & technology.

websites and stores that load fast, look right and convert. built on stacks that still make sense in two years.

a website is the hardest-working asset in most smb's - it sells, it explains, it captures leads, it represents the brand at every hour. we build sites and platforms that do that job well, and we build them to be maintainable long after launch.

web design & build

brochure sites, lead-gen sites, complex multi-page builds. design and development handled together, so what gets designed actually gets built. we work in concretecms, wordpress and headless setups - whichever fits the brief, the budget and the team who'll run it after we hand over.

interactive tools & calculators

sometimes the best lead-gen isn't a form - it's a tool people actually want to use. we design and build calculators, configurators and product finders that answer a real question, then hand the visitor to the right product.

see them working. our glasses frame size & fit calculator turns buying glasses online from guesswork into millimetres, draws your frame to scale on your own photo, and matches you with independent makers. the maths bridge started as something we built for our own kid, and turned into a thirty-day practice app that adapts to the student and shows its working on every question. strumble. started as a song-to-tab converter and grew into a full guitar companion: it turns a recording into chords and tab, reads a tab or chord sheet from a text file, a pdf or a photo, lets a student write their own song note by note, and keeps a chord library, every scale on the neck, a tuner, a metronome and a practice mode in the same page. the same engine runs a product finder or an onboarding wizard.

all three built the way we build for clients - fast, self-contained, no plugins to break.

strumble. a product, not a demo

strumble. is what a small, well-made web app can be. one page, no install, no account, and a guitar student can:

  • capture a song. import a file (or a handful at once) or play it in, trim the part that matters, and get the chords, the strumming and the tab, in time with the recording. every bar is editable, because a machine reading a recording will sometimes be wrong.
  • write their own. a tab editor built for a finger on a phone: tap a string to write a note, pick the fret, finger and length, add hammer-ons, slides and bends, drag notes about, hold to delete. chords are picked from the sheet and dropped in a bar each. a lead can be recorded straight off the fretboard.
  • bring in a sheet. ascii tab, chords over lyrics, a pdf, a photo or screenshot of a page, or a midi file becomes an editable song, with the timing and tempo set by the student where the sheet doesn't say.
  • look anything up. every chord as a playable card with its shape, every scale in every key laid out on the neck, with the chords that fit under it.
  • practise. play, loop bars, slow it down without changing the pitch, drill the chord changes, keep time with the metronome, tune up. a line-delay calibration taps along to eight clicks and sets itself for whatever speaker or headphones are in use.
  • learn it by doing it. built-in live tutorials, one per task: the tour says what to tap, the student taps it, and it moves on. no next button to confuse the lesson.
  • take it away. a printable song sheet with shapes, bars, strum and tab, a share link that opens on any device, and a guitar-and-pedals sound engine that plays the whole thing back as an acoustic or an electric.

it runs at its own address, strumble.app, as one self-contained page, sends its own analytics events into matomo and gtm, and has shipped more than seven hundred and fifty versions through an automated build and deploy, each one checked by scripted browser tests on desktop and phone before it goes live. that is the discipline we bring to a client's tool too.

machine learning in the browser

some problems need a model, not a formula. strumble. runs a neural note-detection model, a key and tempo estimator, a chord engine and a text reader for photos of tab sheets entirely on the visitor's own device. the audio never leaves the browser, there is no server to pay for, and it keeps working on a school laptop with the wifi switched off after the first load. the guitar it plays back with is built on the device as well: a small pack of real acoustic guitar recordings, about the size of one photo, shaped by its own strings, pickups and pedals, with a fully synthesised voice standing in until the recordings arrive.

that suits anything where the data is sensitive, the volume is unpredictable, or a per-request api bill would kill the idea before it starts. it also sets an honest bar: we measure these tools against real recordings with human annotations rather than against a demo, keep a benchmark of riffs and whole songs that every engine change is scored against, and we say plainly what they get wrong. every bar strumble. produces is editable, because a machine reading a recording will sometimes be wrong and the tool should expect that rather than hide it.

ecommerce & shopify

shopify and shopify plus stores designed to convert. theme builds, custom development, app integration, checkout optimisation. we set up stores that don't fall over at scale and don't trap you in a setup that needs a developer for every product launch.

cms & platforms

content management built around the people actually using it. concretecms, wordpress, custom builds. permission setups, editor workflows, multilingual content - the bits that get ignored on most builds and cause the most pain six months in.

tech integration

making the stack talk to itself. crm to website, ecommerce to email, analytics to ads. api's, webhooks, middleware - the unglamorous work that turns separate tools into one system. less swivel-chair admin, more time spent on actual work.

why ink.

we've built and maintained sites and stores for two decades. we know what breaks at scale, what slows things down, and what sneaky vendor lock-in looks like before you sign. the build is the easy part - we build for the two years after launch as well.

faqs

which cms do you recommend?

depends on the business. concretecms for content-heavy sites with editor workflows, shopify for ecommerce, headless setups for product-led businesses with engineering resource. we recommend based on the use case, not the platform we're keenest to sell.

can you migrate our existing site?

yes. we've migrated sites across cms's, hosting and domains. we plan migrations to protect seo, redirects and historical content.

do you offer ongoing maintenance?

yes. support, updates, security patching, plus iterative improvements after launch.

can you build a calculator for our products?

yes. anything that helps a customer answer 'which one fits me?' - sizing, configuration, product matching. we can white-label our frame fit calculator for eyewear retailers, or build one from scratch around your catalogue. self-contained embeds that drop into whatever cms you already run.

can you build ai into our site without sending our data to openai?

often, yes. models small enough to run in the browser already handle a lot: audio, images, classification, search. strumble. does real audio transcription that way, with no server at all. where a hosted model genuinely is the right answer we'll say so, and we'll tell you what it costs per request before you commit to it.

can you build something like strumble. for our industry?

yes. the parts are general: audio and image models that run on the device, an editor that works with a finger on a phone, embedded tutorials, printable output, share links, analytics wired into whatever you already run. a music school, a clinic, a trade supplier - the domain changes, the way we build it doesn't. bring us the question your customers keep asking and we'll tell you whether a tool is the honest answer.

how long does a website project take?

a brochure site can be 4-6 weeks. a complex build with custom development is typically 12-16 weeks. ecommerce sits in between.