Does Attio have a bulk merge?

No. Attio's interface merges two records at a time, and its REST API merge endpoint also takes two records per call. There is no built-in way to select many duplicate groups and merge them in one action.

How do you merge hundreds of duplicates in Attio?

Either script the two-record merge endpoint yourself — building duplicate groups, ordering each group into a chain of pairwise merges, and handling the new record id each merge returns — or use a tool that does this for you. Deduptio scans an object, groups the duplicates, and merges a reviewed set in one run, chaining native pairwise merges per group.

Is bulk merging in Attio safe?

The merge itself is Attio's own and loses nothing, but it cannot be undone from Attio: the losing record is deleted permanently. Bulk merging is therefore only as safe as the rule that produced the groups. Merge an exact-email batch in bulk with confidence; review fuzzy or name-based matches individually before including them.

How long does a bulk merge take?

It is bounded by Attio's API rather than by the tool. Each merge is a separate call, some of which Attio processes asynchronously, so a few hundred groups take minutes rather than seconds. Deduptio persists progress per group so an interrupted run resumes instead of restarting.

How to bulk merge duplicates in Attio

Last checked September 1, 2026 · Deduptio is our product. Every claim about another tool comes from its public site or documentation and is linked in place — tell us if something is out of date.

The short answer

Attio has no bulk merge. The UI merges two records at a time and the API's merge endpoint takes two records per call. Clearing a real backlog means either scripting the pairwise chain yourself or using a tool that does — and in both cases the hard part is producing groups you trust enough to merge without opening each one.

Why there is no bulk merge

A merge is a destructive, irreversible operation: one record survives, the other is deleted and cannot be recovered. A bulk version of that needs three things Attio deliberately does not ship — a way to define what counts as a duplicate, a review surface, and a way back. Without all three, a bulk merge button is a footgun with a progress bar.

So the interface offers the safe primitive: two records, chosen by a person, one merge. The API mirrors it.

The four ways people actually do this

Approach	Realistic for	What it costs you
Merge by hand in Attio	Up to roughly 30 groups	A morning. Free, and nothing to set up
Export to CSV, find groups, merge by hand	Sizing the problem, not fixing it	You still merge every pair manually — see why the spreadsheet stops here
Script the merge endpoint	Engineering teams with a clear rule	Grouping, chain ordering, async handling, rate limits, and no undo if the rule was wrong
A dedupe tool	Hundreds or thousands of groups	A subscription, and rules you have to think about once

If you script it: the parts that bite

Nothing here is exotic, but each one is a bug you only find in production:

Grouping is not pairing. Three records sharing an email are one group, not three pairs. Merge them as a chain: A into B, then the result into C. Merging the three pairs independently means passing a deleted record to the second call.
Each merge returns a new record. The id you got back is what the next hop must use. Persist it per hop, or an interrupted run resumes with a stale id.
Some merges are asynchronous. A call can be accepted and finish later. Treat an accepted-but-unfinished merge as committed and poll — retrying it blindly is how you double-merge.
Uniqueness constraints fight you.You cannot write a losing record's email onto the survivor while the loser still holds it. Field fixes belong after the merge, not before.
Lists end up with two entries. If both records were in the same list, the survivor now sits in it twice. See duplicate list entries.
There is no undo. If the rule was wrong, a thousand merges are a thousand deleted records. Snapshot both records before each merge, or accept that the run is final.

The order that makes bulk merging safe

Split by confidence, not by volume. Run the exact-identifier rule first — same email, same domain, same external id. Those groups are safe to merge in bulk. Fuzzy and name-based matches are a second, separate pass with individual review.
Dry-run the whole batch. A preview that writes nothing tells you which record would survive in each group and what would move. Read a sample of twenty.
Merge a small batch first. Ten groups, then open them in Attio and check activity, references and list membership before running the rest.
Then run the bulk pass, and keep the restore window in mind while you spot-check — an error caught the same day is recoverable, one caught next week is not.
Schedule a scan so the backlog does not rebuild while you are not looking.

How Deduptio does it

A scan pages the object, applies your rule, and produces groups with the reason each one matched. In the review queue you can merge a single group, skip the ones that are not real duplicates — skipped groups do not come back — or filter and select the set you trust and merge it in one run. Each group is merged as a chain of Attio-native pairwise merges, with per-hop progress persisted so an interrupted run resumes rather than restarting. Every merge is snapshotted first and restorable for 24 hours.

See what is duplicated before changing anything

Connect your Attio workspace, write one match rule, and run a scan. Scans are read-only — you see every group and the reason it matched before anything is merged.

Start a free Attio duplicate scan

How to merge records in Attio — the manual version, and why it is fine for small numbers.
Attio duplicate detection with the API — the endpoints involved if you are building this yourself.
The Attio deduplication guide · Compare tools