Person matching and address intelligence

Linksync

Match people across CRM lists and imports, clean messy addresses, resolve them to real places, and return locality-aware details your data workflows can trust.

Try the demo View API docs Match people How locality works
P

Match people

Use Splink to compare primary CRM records with imports, API records, or CRM-supplied candidate blocks.

N

Normalize addresses

Parse free text or structured fields into consistent address components for downstream systems.

E

Enrich places

Resolve addresses against a verified local database — the federal National Address Database, state address authorities, and OpenAddresses as a fallback — with coordinates, locality details, and postcode checks.

PII control

A clear lifecycle for sensitive data.

Linksync can hold matching records in the cloud when you want managed batch processing, but it also supports runtime matching where your system sends candidate blocks only when a match is requested.

1

Cloud-held matching records

Upload person and address files, train matching behavior, run matching jobs, and download results. When the job is done, users can remove source data, results, and matching inputs while retaining only training insight.

2

Runtime candidate blocks

If you do not want PII persisted, your CRM can send candidate records and block ids at runtime. These are matching cohorts, such as records grouped by postcode or last-name initial. Linksync scores those candidates without needing to store the underlying person file.

3

Encrypted object storage

Uploaded files and generated outputs are stored in AWS S3-backed object storage, isolated by account paths and encrypted at rest by S3 server-side encryption.

4

Decoupled processing

The web application handles authentication and workflow control, while background workers process address and matching jobs separately from the public request path.

You can upload training records to teach the matcher, then purge the person and address data. Linksync issues deletion calls for the underlying S3 objects and keeps only the matching model settings and training analysis, preserving useful matching insight without retaining the original PII.
Live demo

Try address normalization

Linksync can take a full address in one box, individual structured fields, or a zip code first followed by live typeahead.

When the match is ambiguous (e.g. a building with multiple units on file), Linksync can return other plausible matches alongside the best one.
Uses whichever of city/state/postcode you've already filled in to scope live suggestions -- no separate address box needed.
Fill in whichever fields you have -- housenumber/street plus at least one of city, state, or postcode is enough.
Both are optional -- a zip code resolves its state automatically and restricts the State field above to just the state(s) that zip actually covers (e.g. only MA for 02459). Leave both blank and just include the city/zip in the address text below instead.
Type any part of the address (house number and/or street, in any order) -- suggestions update as you type. If you skipped Zip/State above, include the city and/or zip here too (e.g. "47 Glen Avenue, Newton, MA 02459").

Response preview

The normalized JSON response appears here after you run the demo.

Results include parsed fields such as street, housenumber, city, state, postcode, coordinates, and locality details when available. Learn how locality is chosen.