Fewer repeated hours. Faster answers. Days that hold together.
8lax builds and runs its own SaaS. The products sit in markets that have nothing to do with each other and do one job between them: take the work a person should not be repeating, and hand back the hour it was eating.
For a business that means a lower cost of getting to yes, a proposal out the same day instead of next week, and an audit that does not consume a quarter. For someone whose job is looking after other people it means a day that holds together. For everyone else it means their own life stopping asking twice.
We use AI where its output can be checked and nowhere else. That is a narrower promise than the one you are used to reading on pages like this, and a more useful one.
What the work is for
Four markets, one question underneath all of them: which hour is this giving somebody back, and can they check that it did.
Work a business pays for twice
An access review assembled by hand every quarter, from the same screens, by someone whose job is not assembling access reviews. A scope written from scratch for a client who has described roughly the project you did three times already. The work repeats, so it should be automated, and what comes back has to be checkable, because a number nobody is allowed to question is not an answer. Attestly, ScopeBench.
The client who is not waiting
Most work is not lost on price. It is lost on the week between the request and the reply, while somebody else answers first. Getting a real scope and a real staffing plan back the same day is worth more than shaving a percentage off the rate, and it is the difference between winning the work and being the third quote in the folder. ScopeBench.
A day other people depend on
Care is scheduling with consequences: the medication at the hour it is due, the visit that actually happened, the handover that currently lives in a message thread. It is also the least automated work there is, and much of it is done unpaid by someone's daughter. Under our ehelpx.com domain. CarePlanner.
Life outside the job
LiCly keeps the list of everything that comes round again: the service, the filter, the check-up, the renewal. None of it has to live in your head. TopBin9 is for the part of the week nobody has to justify. Not every product has to be about efficiency, and pretending it is would be exactly the stretch this page is trying to avoid. LiCly, TopBin9.
Work
Five products, all our own. Ordered by how far along they are, not by how much we like them.
TopBin9
A football media portal. Today's fixtures and the latest results across Europe's major leagues, a running feed of news and transfers, and insider posts. All of it in one place, instead of five tabs and two apps.
The parts people stay for are the ones that are not a feed: a match simulator that plays out any fixture so you can argue about the result before it happens, Guess the XI before kick-off, and a leaderboard that keeps score of who was right.
It is the only product here anybody can open right now without asking us for anything.
Attestly
Jira tells you who has access today. It does not tell you who had access on 31 March, or who confirmed that the list was correct. The organisation audit log keeps events for 180 days, so the evidence of a first-quarter review is gone before a third-quarter audit asks for it.
Attestly takes a snapshot, assigns a named reviewer, and makes them sign off each line with a written reason. It shows what changed since the last review, and exports an evidence pack that verifies on its own, on the auditor's machine, without our software installed.
CarePlanner
Care that keeps to the schedule. A plan for the day: the medication at the hour it is actually due, the blood-pressure reading, the walk, the glass of water. Reminders that fire, and a view of what has been done rather than what was meant to be.
It is built first for the people who do this for a living: self-employed carers with several clients, home-visiting services, care and nursing homes with a rota to hold together and a handover that currently happens by message. The same plan works, unchanged, for one person looking after a parent.
In closed beta on our own environment. There is no public address yet, and we would rather say that than point you at a staging server.
ScopeBench
Paste the client's email. Get back a scope, an estimate, and the names of the people on your own bench who would do it.
Pre-sales in an agency is two jobs pretending to be one. Work out what the client is actually asking for, and work out who is free to do it and when. The tools that know your team have nothing for the first job; the tools with AI for the first job know nothing about your team. Answering both at once is the whole point, because it is the combination that decides whether you can say yes.
Every estimate comes back as a range, never a single number, with the assumptions behind it printed on the proposal where the client can see them. Work nobody on your bench can cover is shown as a gap rather than quietly absorbed into someone's week. It is a draft for your delivery lead to argue with, not a figure to hide behind.
Built and working, and not open yet. If this is your Monday every Monday, write and we will talk before it is.
LiCly
Everything you own has a next time. The oil change, the water filter, the air-conditioner filters, the dental check, the descaling, the subscription that renews whether or not you meant it to. LiCly keeps the list and tells you what is due.
Cycles run on time or on use, whether that is every six months, every 10,000 km or every so many hours. Each item keeps its own history: when it was done, what it cost, a photo, and who did it. A list can be shared, so a household closes items together instead of one person being the one who remembers. Presets get a real list started in a couple of minutes rather than leaving you with an empty screen.
It runs in the browser and keeps its data there. A business version is in research: we are checking the hypotheses against the market before building anything, which is slower than building and considerably cheaper than being wrong.
How we build
Four habits that show up in the code, not just on this page.
Claims are testable
If the software says it never sends your data anywhere, a test refuses to build when someone adds an outbound call. If it says logs contain no personal data, a test reads every log statement and fails on the one that would print a name. Promises that only live in documentation drift.
AI only where you can check it
ScopeBench uses a model to draft a scope, and every number it returns is a range with the assumption printed next to it, and it still works with the model switched off, saying so on the draft. The Forge apps use no model at all, and the privacy policy says that in as many words. The difference is a decision, not an accident.
A person stays responsible
Attestly will not certify anyone's access; it makes a named reviewer do it and write why. ScopeBench produces a draft for your delivery lead to argue with, not a figure to obey. CarePlanner records what somebody actually did, not what the rota assumed. Software that quietly makes the decision is easier to build and worse to answer for.
We say what we cannot do
The Forge apps are read-only: they can show you that someone should lose access, and they cannot take it away. Tamper-evident is not immutable, and we write that on the report itself rather than letting the word do work it has not earned.
Getting in touch
Something that is running
TopBin9 is open now, and Attestly is running with its Marketplace listing in preparation. Bug reports and feature requests reach the people who wrote the code rather than a support tier, and the founder reads them.
Something not open yet
CarePlanner and ScopeBench are being used by a small number of people while they are finished. Say which one and what you would use it for, and we will tell you honestly where it stands and whether it is ready for you.