Extract & Transform
Raw customer data is messy — different casing, formats, and field names in every system. Tilores ingests records from any source and turns them into clean, normalized, enrichable entities, ready to be matched.
submit()Watch a record get cleaned up
Step through the pipeline to see a single messy record become a canonical, match-ready entity. Changed fields are highlighted at each step.
More than a cleanup — a foundation for resolution
Any schema, no migration
Push records in whatever shape they already have. Tilores maps them to a flexible entity model — you never reshape your data warehouse to fit ours.
Normalized for matching
Names are cased and split, addresses parsed and expanded, phones formatted to E.164, dates made ISO. Clean inputs mean the matching engine compares like with like.
Enriched on the way in
Derive country codes, split name parts, canonicalize email providers, and attach your own computed attributes — before a single edge is drawn.
One call to ingest.
Preview before you commit.
Send records with submit. Records pushed together are linked with a STATIC edge, so data you already know belongs together stays together.
Not sure how a batch will resolve? Use submitWithPreview with dryRun to see the resulting entities without writing anything.
mutation IngestCustomers {
submitWithPreview(
input: {
records: [
{ id: "crm-8842", name: "Jon Smith",
email: "jsmith@googlemail.com" }
]
dryRun: true
}
) {
entities { id score }
duplicates
}
}See scattered records assemble into one entity.
Try it with your own data
Download the evaluation build and run the full ingest pipeline locally — up to 100,000 records, no signup.