The software we built once and use on every project, and the overnight machine that checks our code.
The Agent Server is our autonomous agent server, and it works through the night. It runs AI agents that look for problems, reproduce them, fix them and check each other's work. Fixes that pass every test are set aside for release. Nothing goes live until one of us has read it.
At 22:00 the Agent Server wakes eight agents, each with one job.
They sweep the code and flag what looks wrong.
A tester reproduces the problem, a builder fixes it.
A verifier re-runs the work. Not everything passes.
Fixes that pass every test are set aside for release. Nothing goes live until one of us has read it.
Nothing is deployed until a person has reviewed it.
At 22:00 the Agent Server wakes eight agents, each with one job: scout, tester, auditor, backend, builder, security, upgrader and toolsmith.
They sweep the code and flag what looks wrong, like a form value that is not checked or money rounded the wrong way.
A tester reproduces the problem, a builder fixes it.
A verifier re-runs the work and has to agree. Not everything passes.
Fixes that pass every test are set aside for release. Nothing goes live until one of us has read it.
Nothing is deployed until a person has reviewed it.
These are the parts we built once and use on every project: sign-in, admin screens, records of who changed what, page design and documents. When one of them gets better, every app built on it does too.
Sign-in with a second check, roles, admin screens and a record of every change come with every app on the first day.
A library of designed, tested page sections. A new brand needs a change of colour and type, and every section follows.
Pages are put together from Lattice sections, edited in a visual studio and published in one step. Your team changes the words without calling us. This site runs on it.
A page is composed from tested sections, then edited by your own team.
Proposals, reports and statements generated straight from the data, in your house style. The figures in the document are the figures in the system.
Short animations that explain a process or a product, timed so every line can be read.
Our autonomous Agent Server runs AI agents that test, review and upgrade our code. Fixes that pass every test are set aside for release, and a person reads them before anything goes live.
It starts with an assessment.