MFG Calcs Answer API
The Answer API from MFG Calcs — 1 operation(s) for answer.
The Answer API from MFG Calcs — 1 operation(s) for answer.
Every API here is available over the APIs.io API and to AI agents over MCP.
One button, every client — Claude, Cursor, VS Code and the rest.
https://apis.io/mcp
find_apisBrowse and filter every API in the catalog.get_api_artifactsOne API's artifacts, grouped by type.get_openapiThe primary OpenAPI for this API.find_similar_apisAPIs that look like this one.apis_io_searchSTART HERE — APIs, providers and tags for one query, each with its total.resolveTurn a domain, URL or GitHub org into the provider it belongs to.find_cohortsEvery scored population of providers in the catalog.curl "https://apis.io/api/v1/apis/mfgcalcs-com:mfgcalcs-com-answer-api"
curl "https://apis.io/api/v1/apis?limit=25"
Calls without X-API-Key need no key. Calls with it need a paid plan — get a key.
Free tier, no form to fill in. Signing in shares your email address with us — we store it to create your key and to recognise you if you sign in with another provider. By signing up you are subscribing to the APIs.io and API Evangelist newsletters — you can unsubscribe from any issue. See our Privacy Policy and Terms.
A second provider on the same verified email joins the account you already have.
openapi: 3.2.0
info:
title: MFG Calcs Manufacturing Data and Calculation Answer API
version: 1.0.0
description: 'The whole mfgcalcs.com platform over HTTP, keyless. 34 data endpoints serve live U.S. manufacturing intelligence: the effective calculated-duty index and estimate, per-family effective-rate gap and statutory duty-free line coverage, sourcing shifts and concentration, customs unit-value movers, the cost-pressure index and its diffusion, manufacturing PMI, state cost indices, the wage board, six-month forecasts with their own accuracy ledger, reshoring readiness and indexed break-even comparisons, and an exact linear-programming sourcing optimizer. Alongside them: 5,384 deterministic manufacturing calculators, one search across everything the site knows, and an answer endpoint that returns a scoped, source-dated answer or a structured clarification request for a natural-language question.
Everything is computed from committed public data (USITC, BLS, EIA, Census, BEA) rather than a model, so identical inputs always return identical output. The trust surfaces are part of the API, not a footnote: /api/vintages recovers what any series said on a past date, /api/revisions records every restatement, /api/provenance publishes artifact hashes, and /api/receipts carries Ed25519 signatures. Free with attribution (CC BY 4.0) and no key, ever. An MCP server exposing the same engines as tools for AI agents is at https://www.mfgcalcs.com/api/mcp.
VERSIONING: every path below is also served under /api/v1/ — e.g. https://www.mfgcalcs.com/api/v1/manufacturing-pmi. Prefer the /api/v1 form in code you intend to keep: v1''s response shapes are frozen (fields keep their name, type and meaning; new fields may be added, so ignore unknown ones), and any breaking change ships as /api/v2 instead of altering v1. The unversioned paths are permanent aliases and will not be removed. Responses carry an X-API-Version header. See https://www.mfgcalcs.com/developers and docs/methodology/API_VERSIONING.md.'
contact:
url: https://www.mfgcalcs.com/contact-us
servers:
- url: https://www.mfgcalcs.com
tags:
- name: Answer
paths:
/api/answer:
get:
operationId: answerQuestion
summary: Answer manufacturing questions, collect calculator inputs, or clarify a subject
description: 'The deterministic answer engine shared by MCP ask, MFGSearch, and MFGChat. No LLM. Statistical answers identify their period, frequency, scope, and measure. An unresolved subject or tied product family returns resolution=needs-clarification, stat=null, and no citation. A selected calculator collects required inputs, validates supplied values, and executes when complete; missing or invalid inputs request clarification without claiming a calculated result. Calculator guidance may include field counts, accepted values, and a link to the published method. Pass the latest opaque context unchanged to retain the confirmed subject, country, period, and calculator inputs across replies and corrections. Documented defaults are used by guided calculations only for an explicitly requested illustrative example. The detail projection controls depth: answer is compact; evidence adds scoped evidence, calculator guidance, and relevant adjacent facts; full is the compatibility default and preserves the forensic payload. Try ''effective tariff on bearings'', ''what is OEE'', or ''Calculate OEE''. Direct structured calculations are also available at /api/calc; broader entity search is at /api/search.'
parameters:
- name: q
in: query
required: true
schema:
type: string
maxLength: 500
description: A full question in plain English
- name: context
in: query
required: false
schema:
type: string
maxLength: 1024
description: The opaque context returned by the preceding answer. Pass the latest token unchanged on the next turn to retain the confirmed entity, country, period, intent, and calculator inputs or resolve a subject choice. Calculator drafts are revalidated against the current field schema. Context expires after 30 minutes; malformed or expired context is ignored. It is caller-held continuation state, not authentication or evidence that supplied measurements are correct.
- name: detail
in: query
required: false
schema:
type: string
enum:
- answer
- evidence
- full
default: full
description: Response depth. answer returns the sentence, headline statistic, citation, trust blocks, and context while omitting points, adjacent facts, and the tool payload. evidence adds a bounded relevant series window, scoped evidence, calculator guidance when applicable, and up to five adjacent facts. full returns the compatibility-default forensic payload.
responses:
'200':
description: A composed answer, calculator guidance, or structured clarification. Branch on `outcome` (additive) and `fulfilled` before treating a reply as complete; only answered fulfills a calculation, and a related-resource reply fulfills a request for that link. `resolution` stays as the frozen legacy signal and cannot express unsupported or related-resource results, which carry fallback=true as a compatibility bridge. Unresolved subjects have stat=null and no citation. A selected calculator requesting inputs also has stat=null (outcome=needs-input), but may name supplied values and link its published method; that link is not a completed result. Input collection can have fallback=false, so neither HTTP success nor fallback alone establishes completion. In evidence/full detail, inspect payload.calculatorGuidance for accepted inputs, missing fields, errors, and guidance kind. Unsupported requested dimensions may be identified by `unsupportedAttribute`.
content:
application/json:
schema:
type: object
required:
- query
- intent
- resolution
- detail
- answer
- confidence
- stat
- url
- source
- receipt
- freshness
- context
properties:
query:
type: string
intent:
type: string
resolution:
type: string
enum:
- resolved
- needs-clarification
description: 'Frozen legacy compatibility signal: whether the response answers the question or requests a subject, product-family choice, missing calculator inputs, or corrected values. It cannot express unsupported or related-resource results; branch on `outcome` instead. New enum values will never be added here.'
outcome:
type: string
enum:
- answered
- needs-input
- needs-clarification
- invalid-input
- unsupported
- related-resource
description: 'Authoritative fulfillment outcome (additive). answered: The requested operation and scope were satisfied; a scoped statistic or calculator result is present, and fulfilled is true. needs-input: A calculator was selected but declared inputs are missing; values already supplied are retained and stat is null. needs-clarification: The subject or product family is ambiguous, or a rejected value needs a replacement; no number is guessed and no citation is returned. invalid-input: Supplied values were rejected, for example outside a documented bound or in conflict with another field. unsupported: The question named a dimension this platform has no data for; unsupportedAttribute names it, and related published facts may still be included. related-resource: A relevant link stands in for the request, such as a definition, guide, or dataset. Legacy cases the resolution enum cannot express carry fallback=true as a compatibility bridge; unmodified resolution-only clients cannot distinguish unsupported from related-resource.'
fulfilled:
type: boolean
description: Additive completion flag. True only for answered, and for a related-resource reply that fulfilled a request for that link. Never true for needs-input, needs-clarification, invalid-input or unsupported. Legacy consumers that only read `resolution`/`fallback` must continue to treat fallback=true as not fulfilled.
detail:
type: string
enum:
- answer
- evidence
- full
default: full
description: The response-depth projection applied to this result.
answer:
type: string
description: The composed answer or clarification. Calculator guidance may list accepted inputs and missing fields without claiming a calculated result.
confidence:
type: string
enum:
- high
- low
description: Confidence in the selected subject or answer. A known calculator can have high confidence while still requesting required inputs; this is not a completion or universal accuracy score.
stat:
oneOf:
- type: object
required:
- value
- unit
- period
- frequency
- scope
- measure
- change
properties:
value:
type: string
unit:
type:
- string
- 'null'
period:
type: string
frequency:
type: string
scope:
type: string
measure:
type: string
change:
type:
- string
- 'null'
changeReadout:
type: object
description: Optional server-composed display and numeric change metadata for the shared answer cards.
now:
type: object
required:
- display
- periodLabel
- bridge
properties:
display:
type: string
periodLabel:
type: string
bridge:
type:
- object
- 'null'
description: Optional current observation and then-to-now change for a historical answer.
invert:
type: boolean
- type: 'null'
description: Headline statistic with its observation period, data frequency, subject/geographic scope, and measure definition. Null for clarification responses and answers without a statistic.
url:
type:
- string
- 'null'
description: Canonical source URL. A calculator collecting inputs may link its documented method; that link is not a completed calculation. Unresolved subjects have no citation.
source:
type:
- string
- 'null'
description: Source label for the answer or published calculator method; unresolved subjects have no source label.
points:
type:
- array
- 'null'
items:
type: object
required:
- period
- value
properties:
period:
type: string
value:
type: number
description: Underlying series when selected by detail. Omitted as null in answer mode; bounded to a relevant window in evidence mode; lossless in full mode.
receipt:
type: object
description: Receipt status and signed claim when available.
freshness:
type: object
description: Measured-cadence freshness state for the evidence actually used, or the reason it cannot be assessed.
context:
type:
- string
- 'null'
description: Opaque continuation token to pass unchanged with the next question. Product clarification retains country, period, intent, and candidate choices without a guessed product. Calculator turns retain the selected tool and validated draft inputs, including corrections, within the existing context limits.
alsoMatters:
type: array
items:
type: object
required:
- period
- frequency
- scope
- measure
properties:
period:
type: string
frequency:
type: string
scope:
type: string
measure:
type: string
description: Relevant adjacent facts selected by evidence/full detail; each numeric fact carries its own period, frequency, scope, and measure.
payload:
description: Source-tool payload. Null in answer mode, scoped in evidence mode, and lossless in full mode. Calculator responses include calculatorGuidance; completed calculations also retain the actual calculator results and derivation. Guidance or an illustrative example must not be treated as execution of the caller's complete measurements.
properties:
calculatorGuidance:
type: object
properties:
kind:
type: string
enum:
- guidance
- inputs
- calculation
- example
- unresolved
acceptedInputs:
type: object
additionalProperties:
type: number
missingInputs:
type: array
items:
type: object
required:
- key
- label
- unit
properties:
key:
type: string
label:
type: string
unit:
type: string
errors:
type: array
items:
type: string
metadata:
type: object
followups:
type: array
items:
type: string
description: Supported next questions. Use these when an input clarification has no subject choices, passing the latest context with the selected question.
sourceTool:
type:
- object
- 'null'
properties:
tool:
type: string
arguments:
type: object
description: Source operation and arguments when available. Calculator guidance identifies get_calculator; an executed calculation or illustrative example identifies run_calculator.
'400':
description: Missing or empty question, an overlong query, or an invalid detail value (code INVALID_DETAIL).
'429':
description: Best-effort function-instance rate limit exceeded; Retry-After is returned.
tags:
- Answer
externalDocs:
url: https://www.mfgcalcs.com/developers
description: Human documentation and MCP setup
x-mfgcalcs-preview:
status: preview-not-published
notice: MCP 2.0 preview documentation. These tools require an explicitly enabled preview deployment; this inventory does not activate them or prove that a hosted endpoint supports them. Read that endpoint's discovery response before calling.
metadata: /mcp-preview.json
schemaReference: /openapi-mcp-preview.json
examples: /mcp-preview-examples.json