Stock & inventory assistant

The LLM bot that never gets numbers
wrong.

Owners and customers ask about stock in plain language — quantities on hand, totals, percentages, what's running low — and every number in the answer is computed by a deterministic math engine, not the model. Zero hallucinated digits, even at scale. That exactness is the moat: general-purpose chatbots guess at numbers, MathBot never does.

> Which is bigger, 9.11 or 9.9?
The largest number is 9.9, the smallest is 9.11. Descending order: 9.9, 9.11.

Why this isn't "just another chatbot"

General-purpose LLMs routinely get simple number comparisons and decimal addition wrong. MathBot prevents that by design.

=

Guaranteed exactness

Numbers are computed with exact decimals and fractions, never binary floating-point rounding errors like 0.1 + 0.2 ≠ 0.3.

Multiple LLM providers

Claude and OpenAI behind the same interface — pick a provider per request, no vendor lock-in.

{ }

Simple REST API

One endpoint, one API key. Send a question in plain language, get back an exact structured answer.

Comparisons, percentages, statistics

Number comparisons, arithmetic, percentages, rounding, number properties (even/odd/prime), sorting.

How it works

The LLM is only ever allowed to understand the question — never to invent a number.

A question in plain language

The user asks something like "what is 15% of 240" or "which is bigger, 3/7 or 0.43".

The LLM recognizes the operation

The model only ever calls a structured tool and extracts the numbers from the text — it never computes anything itself.

The exact engine computes it

A deterministic Python engine (Decimal / sympy) computes the precise result and returns it in the answer.

An API in three lines

Production API keys are issued per license — contact us to get one.

# curl example
curl -X POST https://your-domain.com/api/v1/ask \
  -H "X-API-Key: mb_your_key" \
  -H "Content-Type: application/json" \
  -d '{"question": "Is 17 a prime number?", "provider": "anthropic"}'

# response
{
  "ok": true,
  "provider": "anthropic",
  "operation": "number_properties",
  "text": "17 is odd, prime. Divisors: 1, 17."
}