Services
We teach it, we measure it, we test it.
The work behind the publication, done privately for one organisation. Your hardware, your questions, your report.
Training
Half day or full day, remote, up to 12 people
For a team that wants to run models on its own machines
Your stack, not ours
This is not a tour of how StrideNote works. We prepare it around your hardware, your constraints and the work your team actually does, from a short questionnaire before the day.
People leave able to choose a model for a job, run it on the machines they already own, measure what it costs them in memory and speed, and check what it sends out. When the tools change next year, the method still holds.
We name every tool we know of for a job and say what each is good and bad at. We are not a reseller and we take nothing from any of them.
Every number we teach from was measured on our own test machine, a 48 GB M4 Pro, and published with the method beside it.
By the end of the day, your team can
- Work out what their hardware can actually take, before downloading anything
- Pick a model for a job, and say why that one
- Run it locally and measure the speed and memory themselves
- See what the tool sends from the machine, and switch off what they can
- Know where it breaks: context limits, quantization, two models at once
They keep
- A recording of the session for the people who missed it
- A written playbook for their setup, not a generic one
Half day or full day, remote, up to 12 people. The full day includes the written playbook. Fixed price, quoted in two working days.
And three more
Same method as the publication, delivered privately
Evaluation
"What should we run, on our hardware?"
We test the candidates on machines like yours and give you the numbers: speed, memory, where each one fails, and which we would pick for your work. The raw measurements come with it.
For the person who has to choose
One question, or a comparison across a shortlist. Priced on the number of candidates and the depth of measurement.
Telemetry audit
"What does our software send?"
The tests behind the Local AI Telemetry Report, run on your build. Code read, every connection listed, traffic captured where we can, each finding graded A, B or C.
For the company that ships it
One application, or a stack that ships together. Priced on the number of components and whether a capture is possible. Re-tests after your fixes cost less.
Documentation
"Write our setup guides."
Playbook-style install and setup guides for your product, written from a real install on our test machine, delivered as Markdown or HTML for your docs site.
For a product or docs team
Priced per guide, on the length and how much of the setup we have to work out from scratch.
How we quote
Every engagement is a fixed price, agreed before we start. No hourly billing, no change requests mid-way: if the scope grows, we say so and re-quote.
Tell us what you want to find out and we send the scope and the price within two working days. Half on agreement, half on delivery, invoiced by StrideNote Studio.
Engagements start in the middle three figures. If your budget is smaller, the playbooks and the reports are free and always will be.
We don't compromise standards.
StrideNote.net is the publication. StrideNote Studio does this work. They share a method and a name, and nothing else.
A company that pays us is not written about on StrideNote.net while the work is under way and for twelve months after it ends. If news forces a mention, the piece says the company was a client. Clients do not see drafts of anything we publish, and nothing learned in private work reaches print.
There is no badge and no seal. "Audited by StrideNote" does not exist, and no finding of ours may be quoted in your marketing.
How it works
- Write to usTell us which of the four you want, and what you are trying to find out.
- Scope and priceWe reply within two working days with the scope, the price and a one-page agreement.
- The workA questionnaire before a training day; your build or your question for the rest.
- What you keepThe recording, the playbook, the report and the raw measurements, all yours.
The course is free
A plain-language video course on running AI on your own machine, made with the same local tools it teaches. Free to watch, no sign-up wall. If your team wants it taught live, around your hardware, that is the training above.
Tell us what you need.
We reply within two working days.