HubSpot ⇄ Salesforce · Contacts
Dataferry connects your HubSpot and Salesforce accounts, has you confirm every field, shows the result on sample records before you approve it, then runs a transfer that verifies every Contact it writes.
Self-hosted. Runs next to your own PostgreSQL and Redis.
Chapter I · the crossing
Every transfer follows the same four steps, and the plan you approve is the plan that runs.
Sign in to HubSpot and Salesforce with OAuth. Dataferry checks that each connection can read Contacts before it stores the credentials, encrypted.
Match real Contact fields from both accounts. Split, join, reformat dates and currency, map values or fill a text template. A configured AI worker can suggest mappings for you to accept or change.
Paste a few source records and read the exact values that would be written. Approval is tied to that plan and expires after 15 minutes, so a changed mapping always gets a fresh look.
The run pages through source Contacts, upserts each one by a unique External ID and reads it back before counting it done. Pause, cancel or retry at any time; a retry updates, it never duplicates.
Chapter II · the manifest
Dataferry starts with a common move: Contacts between HubSpot and Salesforce. Each new route gets the same preview, approval and verified write path before it opens.
Ready today
Next on the chart
Correspondence
Contacts, in either direction between HubSpot and Salesforce. Other apps and objects are in the catalog for planning; they are not transferable yet.
Runs are durable. Resume picks up from the last checkpoint, and Contacts that were already written are matched by their External ID instead of being created again.
It applies your confirmed mapping to the samples you supply, without reading or writing either app. When the run starts, Dataferry checks the connections again.
On your own infrastructure: the web app, API and worker alongside PostgreSQL and Redis. Your credentials and records never pass through a shared service.