Deterministic first · AI second

The engineering layer for electric‑vehicle power.

Six engines compute the answer — NEC load calculations, conductor and breaker sizing, charge curves, depot load management, site economics, tariff selection. Twelve Claude instruments then explain it, audit it, and defend it to an inspector. The arithmetic is checkable. The reasoning is on the record.

55 assertions, hand-checked NEC 2023 No sign-up to compute
Load calculator Live
amps calculated
NEC 220.83(B) · appliances at typical nameplate Open in the panel
Speaks the standards
NEC 220.83 / 220.87NEC 310.16NEC 625.42 SAE J1772NACSCCS1 · CCS2 OCPP 1.6J · 2.0.1ISO 15118OCPI
The problem

A $37 billion market decided by arithmetic
almost nobody runs.

Whether a panel can carry a charger is a calculation, not an opinion. So is the conductor size, the charge time, the depot peak, and whether a site ever breaks even. These get guessed at every day, and the guesses cost thousands.

Global EV charging infrastructure market in 2026, compounding 20.3% to $136.4B by 2033
US public charging ports as of June 2026, up 15% year over year
Upper end of a US home Level 2 installation, against a $1,100 floor
Level 2's share of the charging market — where the decision is personal and the money is real
The core

Six engines that compute,
not six prompts that guess.

Every engine is pure arithmetic against published code tables. Same inputs, same answer, every time — no model in the path. Each one ships with unit tests you can run from the repo.

Read the method

Then the AI layer: every engine result can be sent to Claude for a written review — what the number means, what an inspector will question, what to change, and the three things the model would check next. The computation is deterministic; the interpretation is where the model earns its place.

How it works

Compute, interpret, defend.

Enter what you know

Panel rating, floor area, vehicle, dwell window, tariff. A project holds it all and every engine reads from the same set of facts, so nothing is re-typed and nothing drifts out of sync.

The engine computes

Line-by-line arithmetic with the governing code section printed beside each step. Charts render the charge curve, the load profile, the break-even. Nothing is hidden behind a model.

Claude reviews it

The computed result goes to an instrument with a domain-expert brief: what it means, where it is fragile, what an inspector will ask, and what to do next. Export the whole thing as a project report.

The instrument layer

Twelve instruments, each with one job.

Not one chat box with twelve prompts. Each instrument has a fixed role and a specified output shape — a comparison declares its weights before it scores; a scenario model tables every assumption and names the one that breaks it. Output from any instrument can be handed to any other.

Who it's for

Four people, one set of numbers.

Driver

Buying the first charger

  • Will my panel carry it, and at what amperage
  • What the install should actually cost
  • Which utility rate plan to be on
  • Real charge times for my car, not brochure numbers
Installer

Quoting and permitting

  • 220.83 and 220.87 calculations, printable
  • Conductor, breaker, EGC and voltage drop in one pass
  • A defensible answer when the panel is full
  • Project files per customer, not a napkin
Fleet

Sizing a depot

  • Peak kW with and without load management
  • What the demand charge actually costs per month
  • Service size before the utility study
  • Scenario models with every assumption stated
Operator

Underwriting a site

  • Break-even utilisation, computed not asserted
  • Demand charge as a share of revenue
  • Payback before and after incentives
  • A steelman of the case against building it
Pricing

The engines are free. Always.

Arithmetic against a public code table should not be behind a paywall. What you pay for is the interpretation layer, saved work, and everything a team needs.

Free
$0forever
Everything you need to answer your own question once.
  • All six engines, unlimited
  • All twelve instruments
  • Up to three saved projects
  • Share links on any result
  • Project reports
  • Team workspace
Start computing
Most picked
Pro
$29/ month
For installers and consultants who bill for the answer.
  • Everything in Free
  • Unlimited saved projects
  • Printable project reports
  • Cross-instrument handoff history
  • Priority inference
  • Team workspace
Start free trial
Fleet
$249/ month
For depot planners and site developers modelling capital.
  • Everything in Pro
  • Multi-site depot modelling
  • Five seats included
  • Assumption sets shared across the team
  • CSV and JSON export
  • API access, 50k calls a month
Talk to us
Enterprise
Custom
For utilities, CPOs and OEMs deploying under their own brand.
  • White-label deployment
  • Tenant isolation and SSO
  • Make-ready qualification hooks
  • Custom engines and code editions
  • Deployment in your own cloud
  • Named support engineer
Read the brief
Security & trust

Small system. Short answers to hard questions.

The whole deployed surface is a static front end, two edge functions and a managed Postgres. There is very little to attack, and we can describe all of it in one screen.

Your API key never reaches a browserInference runs through a single edge function. The key exists only as a server environment variable.
Row-level security on every rowProjects and sessions are readable only by the account that created them, enforced in Postgres, not in application code.
No password to stealAuthentication is a one-time email link. We never store a password because there is not one.
Compute without an accountEvery engine runs client-side. If you never sign in, your inputs never leave the browser.
Deletion means deletionSign out clears local state. Account deletion cascades every project and session row.
Questions

The ones that actually get asked.

Is this just a chatbot with a charging theme?
No. The six engines contain no model at all — they are arithmetic against NEC Table 310.16, Chapter 9 Table 8, 220.83, 220.87, 240.4(D) and 250.122, plus standard energy and demand-charge maths. Run npm test in the repo and 55 assertions check the output against hand-worked examples. Claude sits on top to interpret, critique and write up what the engine produced. That order matters: a model that guesses a conductor size is a liability, and a model that explains a computed one is genuinely useful.
Can I use the output for a permit?
Use it to prepare and to check. Charging produces a load calculation in the same shape an inspector expects, with the code section printed beside every line, which is exactly what makes it easy to verify. But it is not a stamped design. Electrical work must be designed and performed by a licensed electrician to the code edition and local amendments in force in your jurisdiction. We say so in the product, not only in the terms.
Which code edition do the engines use?
NFPA 70, National Electrical Code, 2023 edition, 75 °C conductor terminations, no more than three current-carrying conductors in a raceway, no ambient derating applied. Local amendments vary and the engine says so. Enterprise deployments can pin a different edition or add jurisdiction-specific amendments.
What happens if the model gets something wrong?
The number does not come from the model, so a model error cannot change a conductor size or a load calculation. It can produce a wrong explanation — which is why every instrument states the governing code section, flags jurisdiction dependence, and shows the arithmetic it is describing so you can check the claim against the computation sitting next to it.
Do I need an account?
Not to compute. Every engine runs entirely in your browser and works signed out — the inputs never leave the page. An account exists so projects follow you across devices, and so reports and team features have somewhere to live. Sign-in is a one-time email link; there is no password.
Can we run this ourselves?
Yes. The whole thing is a static front end plus two edge functions. Enterprise deployments run in your own cloud account against your own Anthropic key and your own Postgres, with your branding and your code amendments. Nothing phones home.
Who is behind it?
Charging is built by Humble Superintelligence, which builds instrument-grade software for domains where a wrong number is expensive. Charging is the first.

Run the calculation.
Then argue with it.

Six engines, twelve instruments, no sign-up to start.