Who builds it
Lemuny built WIPS. Now it builds for you.
Lemuny is Summerfield's AI engineering team. It designs, builds, tests, and documents, and gates its own work before anything ships. It already built WIPS, live today. The same team builds what's below.
See what it's builtWhat it produces
WIPS
What Lemuny already shipped
Lemuny designed WIPS, built it, tested it, and documented it end to end. Each release runs against roughly 940 automated checks before it ships, and Lemuny gates its own work: architecture review, security findings, and a real-device verification pass, before anything goes out. What's live: WIPS on the Microsoft Store and on the web at wips.summerfieldtech.com, offline-first with zero trackers. The Android release is in testing on Google Play. That is the discipline behind every Lemuny build.
Lemuny gets its own home at lemuny.com, coming soon. Until then, WIPS is the closest thing to watching it work: open it.
The scope doc
Signed before we build
Every engagement starts with a scope doc short enough to actually read. It's what gets signed, and what the defect window measures against.
See the anatomy
This is a condensed version of the actual one-page document a client signs before we build anything. The example below is illustrative: a plausible Process Killer engagement for a general contractor, built to show the shape of the document. It is not a record of a real client.
Offer
The Process Killer. $2,500 CAD flat. Delivery in 10 business days, counted from the first business day after the doc is signed and the first payment is received. What that buys you: the price and the clock are both written down before any work starts. Neither moves once both sides sign.
1. Process name
"Quote follow-up emails." What that buys you: named in the client's own words on the call, not relabeled as "the automation" or "the workflow." The document describes their process, not our jargon for it.
2. Trigger
"A quote is marked Sent in the client's quoting tool." What that buys you: one event starts the whole thing. A fuzzy trigger means a fuzzy scope, so this gets pinned down first, before steps or price.
3. Steps, numbered end to end
- Wait 3 days after the quote is marked Sent.
- If there has been no reply, send a follow-up email from the client's own address.
- Wait 4 more days. If there is still no reply, send a second and final follow-up.
- If the customer replies at any point, stop the sequence and mark the quote Replied.
- Log every step in a status the client can see.
What that buys you: every step exists on paper before a single line of the automation exists. No human touches any step unless a step says so, and in this example none do.
4. Systems touched (maximum two)
System A: the client's quoting tool. System B: the client's email inbox. What that buys you: two systems, named, with access granted by the client and revocable by the client at any time. A request for a third system is not a change order. It is a second engagement, and we say so before quoting it.
5. Success criteria
"Every quote with no reply after 3 days gets one follow-up email within the hour. Demonstrated with a test quote on the handoff call." What that buys you: a behavior that either happens or does not, checked live on the call. Not a claim that "it works well."
6. Explicitly out of scope
"No SMS. No website changes. No cleanup of historical quotes. Anything not written in sections 1 to 5 of this document." What that buys you: the sentence that protects both sides. If it is not written above, it is not included, and both sides agreed to that before signing, not after.
7. Changes
Additions after signing are quoted as change orders: the change, the price, the schedule impact, signed before that work starts. What that buys you: scope does not creep quietly. Every addition gets its own price and its own signature, every time.
8. Defect window
14 days from the handoff call, covering the behavior described in sections 2 to 5 of this document. What that buys you: if something in the signed scope breaks in the first 14 days, fixing it is on us. It is not billed as new work.
Signatures
Client, dated. Summerfield Technologies, dated. What that buys you: one page, both names on it. For a Process Killer, that page is the entire contract.
The audit report
What comes before a Vibe-Code Rescue
A written report ranking what's solid, what's broken, and what's insecure, by how badly each one blocks your launch.
See the anatomy
This is what a Vibe-Code Rescue audit report actually contains, condensed to show its shape. Like the scope-doc example above, this is illustrative: a plausible audit of an AI-built booking tool, built to show the anatomy of the report. It is not a record of a real client.
Verdict
"Buildable. Not launch-ready as it stands. 5 issues block launch. 2 are worth fixing later. 1 is a maintenance note." What that buys you: the headline comes before the detail, so the client is not left guessing at the bottom line while reading findings.
Findings
Ranked by whether they block launch, not by how they were found.
- API keys are visible in the frontend bundle, readable by anyone. Security. Blocks launch.
- The booking form lets two people claim the same time slot. Data integrity. Blocks launch.
- The signup form accepts input with no validation. Security. Blocks launch.
- Admin routes are reachable without a login check. Security. Blocks launch.
- Errors show a blank screen with no message to the user. Stability. Blocks launch.
- Dates are formatted two different ways on two different screens. Correctness. Does not block launch.
- Images are missing alt text. Accessibility. Does not block launch.
- Several dependencies are three major versions behind. Maintenance. Does not block launch.
What that buys you: what is solid gets named too, not only what is broken. If a login flow works cleanly or a payment integration is sound, the report says so, so the fix work does not waste time rebuilding what already works.
Top 5 launch blockers
Findings 1 through 5 above, restated as the scope for the rescue itself. What that buys you: the audit becomes the scope doc for the fix. Five named issues, each one fixed and each one retested. Nothing vague carries into the build.
What's not included
Findings 6 through 8 stay in the report, listed and explained, and are quoted separately if the client wants them addressed too. What that buys you: nothing gets buried. You see the full picture either way, and you decide what is worth paying to fix beyond the 5 that block launch.
Retest
Every fixed issue is tested again against the finding that named it. What that buys you: not our word that something is fixed. A rerun that shows it.
Timeline
The audit report lands within 5 business days of us getting access to the code. What that buys you: a date you can plan around before the first finding is even written.
Three ways in
The Process Killer
One manual process, automated · 10 business days
Pick the manual process that eats your week: quote follow-ups, booking reminders, review requests, invoice generation. We automate it end to end, across up to 2 connected tools.
Every build includes a discovery call, a scope document signed before any work starts, QA, a handoff call, and a 14 day defect window. Anything beyond the scope is quoted separately.
We reply the same business day.
The Ship-It Build
A working product, deployed · 3 to 5 weeks
A real product in production, not a demo: up to 5 core screens, accounts, an admin view, and a production deploy. Two revision rounds are built in, and a 30 day defect window after handoff.
Web builds are $6,500 CAD. Mobile builds are $9,500 CAD and add a signed Android build and store listing materials.
We reply the same business day.
Vibe-Code Rescue
For stuck AI prototypes · 5-day audit, then scoped fixes
You prompted your way to a working demo and then it stopped moving. It happens more than anyone admits. We audit the code, hand you a written report, then fix the top 5 launch-blocking issues.
Prefer to start with the audit on its own? That's $1,500 CAD, credited toward the rescue if you proceed.
We reply the same business day.
Who does the work
Lemuny builds it
Lemuny is our AI engineering team. It designs, builds, tests, and documents. Lemuny built WIPS. Now it builds for clients.
WIPS is the receipt
WIPS is live on the Microsoft Store and on the web, and the Android release is in testing on Google Play.
Before you ask
Is this AI-generated code?
Yes. Lemuny, our AI engineering team, designs, builds, tests, and documents the work, and gates its own release: architecture decisions, security findings, and the final build, checked before anything ships. You talk to the Team Lead, not a line-by-line reviewer. Nothing goes out without clearing those gates.
What if it breaks?
Every offer carries a defect window measured against the signed scope: 14 days for the Process Killer, 30 days for the Ship-It Build. If something in scope breaks in that window, we fix it at no charge. A Vibe-Code Rescue goes further: every issue we fix gets retested, so you are not taking our word for it.
Who owns the code?
You do. Ownership of what we build transfers to you on full payment. Until payment clears, Summerfield retains it. That is written into the scope doc you sign, before any work starts.
Start a project
Tell us where it hurts.
A couple of sentences about the problem is plenty. No pitch deck required. We reply to project enquiries the same business day.