cranium

services

brains for hire.

cranium does not only ship its own products. it also takes on outside work: software and websites built to specification, ai applied where it earns its keep, and repetitive workflows automated.

email tyler@cranium-ai.com

what cranium takes on

  • custom software. desktop applications, web applications, and the internal tools nobody sells because every company needs a slightly different one. the products on this site are the portfolio: a windows desktop app with an installer and an update channel, a multi-tenant saas with billing and third-party oauth, a local hardware controller, and a browser plugin that keeps itself up to date.
  • websites. the page you are reading is one. hand-written html and css, no framework and no tracking, quick on a phone and still legible with javascript switched off. product pages, landing pages, documentation and the ordinary marketing site a business needs in order to look like it exists. search engines and link previews are part of the build rather than a job for afterwards: titles and descriptions written for people, structured data, a sitemap robots can actually read, and a preview card that renders when someone pastes the link.
  • ai consulting. where a language model earns its keep, and where it is an expensive way to do something a query already does. practical work: putting a model behind a real task, keeping its costs and its mistakes visible, and building the plumbing around it that decides whether any of it survives contact with your data.
  • workflow automation. the report somebody rebuilds by hand every month, the data retyped between two systems that will not talk to each other, the document produced one at a time from a record that already holds every field it needs. these are usually the cheapest problems to fix and the ones people have stopped noticing.

discretion

client confidentiality is treated as part of the work rather than something bolted onto it. as a matter of practice:

  • project details are not discussed outside the engagement. clients are not named, work is not shown off, and nothing from one engagement is carried into another or onto this site unless a client asks for it.
  • access is kept narrow and short. credentials, keys and data are asked for only where the work genuinely needs them, held no longer than the task requires, and removed once it is finished — not kept on the chance they prove convenient later.
  • background checks, non-disclosure agreements and vendor reviews are welcome. where your organisation requires one before work can begin, cranium expects to go through it, and will say so plainly and early if any part of it cannot be met.

this describes how cranium works rather than terms being offered: anything binding belongs in an agreement signed by both sides, and this page is not one.

getting in touch

write to tyler@cranium-ai.com. the more of this the first email carries, the more useful the first reply is:

  • what you are trying to do, and what it costs you now — hours a week, mistakes that get through, work you turn down.
  • what already exists. the spreadsheet, the database, the airtable base, the system the new thing has to talk to. half of this work is fitting in alongside something that cannot be replaced.
  • the deadline, if there is a real one, and what it is tied to.
  • what you want at the end — something built, handed over and yours to run, or something with a hand kept on it.

you will hear back with a straight answer: what the work would involve, what it would cost, and whether cranium is the right place for it. small, well-defined jobs are as welcome as large ones.