HealthTech
HelenisCare
Care coordination and referral pathways, built to handle patient data safely. Every pathway ships through the same gate as anything else.
gated build
Windows desktop app · 14-day free trial
Built to ship what can’t afford to break.
teemd is a Windows desktop app that gives you a full AI engineering crew — and a review gate nothing ships without — on the AI accounts you already have: Claude Code, Codex or Ollama. Your AI writes the code. A different AI family checks it. You decide.
For founders building on trust: health records, payment data, real people’s information.
Windows 10 and 11 · free for 14 days from sign-up · no card
On a Mac? macOS is the next platform after Windows. Leave your email and we’ll tell you when it’s ready.
One email when macOS ships. No spam. Terms / Privacy / Pricing
By joining, you can review the terms and privacy policy.
Your request
Round VAT to the nearest cent on invoice totals, and add a test for the half-cent case.
Change ready for review
Rounding now happens once, on the total, not per line.
src/invoice/totals.ts+18−4
tests/totals.test.ts+42
- return lines.reduce((s, l) => s + round(l.vat), 0)
+ const vat = lines.reduce((s, l) => s + l.vat, 0)
+ return roundHalfEven(vat, 2)
GateGREEN
✓
Checked by a different AI familyNo blockers found. Ready for you.
Nothing reaches main until you approve.
Approve and landSend back
44%
of AI code generation tasks introduce a risky security vulnerability — and AI now authors roughly half of all committed code.
— Veracode, 2026 GenAI Code Security Report
78
CVEs have been traced to AI coding assistants since tracking began in May 2025 — and researchers call that a strict lower bound.
— Georgia Tech SSLab, Vibe Security Radar (retrieved Aug 2026)
The gate
Advisory suggestions get ignored under deadline. A gate doesn’t.
teemd’s crew reviews every build the same way — and nothing reaches your repo until it clears.
01 — Build
Your AI writes the change
A full AI engineering crew plans and writes it, on the AI accounts you already have.
flagged
02 — Review
A different AI family checks it
Every change is checked for security and correctness, and reviewed by a model from a different family than the one that wrote it.
in review
03 — You decide
You get the final word
The change reaches you checked and explained. Nothing lands in your repository until you approve it.
shipped
Green
The change reaches you checked and explained.
Yellow
Minor issues, shown to you before you decide.
Red
Nothing reaches your repo until it clears.
Built on what we ship
We run teemd on our own products first, including health and payments.
SamKis Labs builds teemd alongside the products it has to work for. Not synthetic benchmarks: real products we’re building to ship, through the same gate.
FinTech
Novis Pay
A working mobile point-of-sale for Caribbean merchants, built through teemd’s gate. Play Store release pending.
gated build
AgriTech
Farmers Exchange
Produce sourcing and supplier coordination — proof the gate holds up on fast-moving, everyday builds too.
gated build
Subscription
Works with the AI accounts you already have
Claude Code, Codex or Ollama — or any mix. Cross-family checks need two model families set up, such as Claude and Codex, or two open-weight families through your own Ollama account. If one is missing, setup tells you. teemd won’t let a model check its own work. Talk to Jen, your coordinator, to start and steer the work — included on every plan.
Claude Code
Codex
Ollama
The crew
A full AI engineering crew, scaled to the job. Not one model doing every job.
Bigger jobs bring in specialists for user experience, interface and backend work; every job is built, then reviewed twice — once by a different model family — so checking isn’t an afterthought bolted onto the same model that wrote the code.
Control
Review-first, evidence-backed
Nothing lands in your repository until you approve it.
Evidence
A record you can browse
Every delivery’s reviews, gate result and your decision are recorded on your machine, and you can browse them in the Audit log, on every plan.
Data
Your code goes where you send it
Your prompts and code go straight from your machine to the providers you choose. teemd’s servers handle only sign-in, licence and billing.
Moving from Cursor?
Keep the subscriptions you already pay for.
Cursor
12 Oct 2026
Cursor’s new plans bill third-party models separately
Cursor introduced The X Subscription on 12 October 2026. On its new Premium plans and above, third-party models such as Claude and GPT are billed at API list prices, separately from the plan. People on an existing Cursor plan keep their included usage until they switch.
12 Nov 2026
OpenAI plans to end Cursor’s access
OpenAI has said it will end Cursor’s direct access to its models. The proposed date is 12 November 2026.
teemd
A flat teemd subscription; model costs separate
teemd is a flat price per person. Your Claude Code, Codex or Ollama subscriptions and any usage on them are paid separately, to those providers. With two model families set up, a different family reviews every change.
Sources: Cursor forum: Introducing The X Subscription (12 Oct 2026) · Cursor help: plan overview · Engadget.
Pricing
Two plans, priced per person. Your models stay yours.
Start with a 14-day free trial from sign-up, no card. Then Solo or Enterprise, monthly, in US dollars. Model usage isn’t included: you use your own Claude Code, Codex or Ollama accounts. There are no delivery allowances.
Trial
Free
14 days · no card
Solo
$40/mo
per person
Enterprise
$80/mo
per person