Inferrex

/ see it work

Watch Inferrex understand an API.

No signup. Pick a provider and watch schema inference, a golden record, and self-healing run on real classified data from the corpus.

Try it now — infer a schema

Pick a well-known provider and watch Inferrex read its schema and classify every field by business meaning, with confidence scores.

app.inferrex.com/connections

Connections

Stripe API

https://api.stripe.com/v1

96% confidence
1 entity · Customer5 fieldsBearervia OpenAPI spec + live probe

Customer

5 fields

96%
  • email
    Emailemail

    person · email

    99%
  • text
    Full namename

    person · full name

    96%
  • phone
    Phonephone

    person · phone

    94%
  • currency
    Currencydefault_currency

    money · currency

    98%
  • datetime
    Createdcreated

    time · created at

    95%

✓ 5 fields classified — no field mapping needed

Build a golden record

See two providers' records resolve into one authoritative entity — duplicates merged, field authority applied.

app.inferrex.com/pipelines/mappings

Field mappings

The same customer lives in every system under different field names. Inferrex maps each to a shared concept and merges them into one record. Where two systems disagree it picks the source of truth from real lineage — when it last changed and how complete it is. Anything it can't recognise, you map by hand below.

Systems

4 fields contested — choose the source of truth

3 systems → Golden customer

Stripe + HubSpot + Legacy CRM reconciled by canonical concept

Source fieldsGolden customerConf.
Stripeemail2d ago · 99%
HubSpotemail3h ago · 98%
Legacy CRMeml1w ago · 90%
email HubSpot · changed 3h ago — freshest copy
98%
StripenameMar 2024 · 94%
Legacy CRMcust_nm1w ago · 88%
full_name Legacy CRM · changed 1w ago — freshest copy
83%
HubSpotfirstname
first_name
95%
HubSpotlastname
last_name
95%
Stripephone2d ago · 61%
HubSpotphone3h ago · 93%
Legacy CRMtel_no1w ago · 66%
phone HubSpot · changed 3h ago — freshest copy
92%
Legacy CRMnino
national_id
93%
HubSpotcompany
company
88%
HubSpotlifecyclestage
lifecycle_stage
90%
Stripedefault_currency
currency
98%
Stripecreatedat signup · 100%
Legacy CRMsignup_tsat signup · 100%
created_at Stripe · changed at signup — freshest copy
95%

Needs your input — 1 unrecognised field

Inferrex couldn't classify these from their values. Map each one to a target field, or add a new field to the record.

Legacy CRMseg_xunrecognised
41%

One canonical vocabulary lets every system map to the same record — even fields you map by hand. Where two systems describe one field, the source of truth is decided from real lineage. One governed customer, with full provenance.

Break it, watch it heal

Rename a field on the source API and watch Inferrex detect the change, re-infer, and repair the mapping — no human in the loop.

app.inferrex.com/sync/health

Self-healing

Stripe API

Customer · monitored for change

Healthy
  1. Stable
  2. Change detected
  3. Re-inferred
  4. Remapped
datetime
Createdcreated

time · created at

97%

Mapping healthy, data flowing.

Check your stack

Tell us your tools and see corpus coverage across the providers you actually run.

app.inferrex.com/intelligence/coverage

Connect a source

Syncing your stack… 0/6

Your stack, synced against the corpus

Each system is read live — schema discovered, fields classified — no spec needed.

SalesforceREST
reading REST schema…
HubSpotREST
queued
XeroREST
queued
StripeREST
queued
ShopifyGraphQL
queued
ZendeskREST
queued

Infer a schema yourself

Paste a sample record or schema — Inferrex reads it and classifies every field by type.