Source

A parsed quote never invents the number a manufacturer didn't send.

Tamats parses a manufacturer's quote — pasted text or an existing inbox message — into structured commercial terms: unit price and currency, MOQ, lead time, incoterm and a validity date, using plain pattern matching rather than a model's guess. A term the parser can't find stays absent and is listed as unparsed rather than defaulted to a plausible-looking number, and every parsed quote sits in a parsed state until a human confirms or rejects it — nothing here treats a parse as a decision. Currency is never converted; a comparison across quotes in different currencies is flagged as such rather than silently normalised.

A manufacturer's quote arrives as a paragraph of prose, and the actual commercial decision needs it as a handful of comparable numbers — price, MOQ, lead time, incoterm. Doing that translation by hand is small, tedious, and exactly the kind of task where a model would rather guess a number that sounds right than say plainly it couldn't find one. Tamats parses deterministically instead, and treats an unparsed field as information, not a gap to fill in.

What you do with it

Deterministic term extraction

Unit price, currency, MOQ, lead time, incoterm and validity date are pulled from the quote text with plain pattern matching — no model call, no invented figure for a term the text doesn't actually state.

Confidence per field, not a single verdict

Each extracted field carries its own confidence — explicit, inferred, or absent — so a currency symbol read as an assumption (a bare dollar sign mapped to USD) is never presented with the same certainty as a term stated outright.

A parsed to confirmed or rejected state

A parsed quote sits in its own state until a human reviews it in the quote UI and moves it to confirmed or rejected — parsing populates the fields; it never counts as the decision.

No currency conversion

Quotes in different currencies are never converted for comparison — a mixed-currency comparison is flagged as such, so a person makes the exchange-rate call rather than trusting a number Tamats picked.

What it computes over

• —The manufacturer's own quote text, pasted or pulled from a connected inbox message
• —Plain pattern matching over that text — no external price data, no model recall

Where it stops

• A field the parser cannot resolve is left absent and listed as unparsed — it is never defaulted, guessed, or filled from a typical value for the category.
• Nothing here persists automatically. A human confirms the parsed fields in the quote UI before they count as the organisation's record of the quote.
• Currency is never converted — a comparison across quotes in different currencies is flagged as mixed and incomparable rather than normalised to one currency for you.

Questions

Will Tamats guess a quote term it can't find in the text?

No — a field the parser can't resolve stays absent and is listed as unparsed, never defaulted to a plausible number. The point of the parser is that a number you didn't check is never presented as if the manufacturer had stated it.

Does parsing a quote automatically save it as confirmed?

No — a parsed quote sits in a parsed state until a person reviews it in the quote UI and moves it to confirmed or rejected. Parsing only populates the fields; confirming them is a separate, deliberate step.

Can I compare quotes priced in different currencies?

You can see them side by side, but Tamats never converts between currencies for you — a comparison spanning more than one currency is flagged as mixed and incomparable, and the exchange-rate judgement stays yours.