The one CRM here whose own workflow can call the API and read the answer.
- Moves: Contacts, Companies, Custom objects
- Needs: An Attio workspace on a plan with automations, and a Reverse Contact API key
Three integrations, and the first thing to know is that one of them is not an integration at all. Attio reads HubSpot once, as a migration, which is a week wasted if you arrived expecting a sync.
A CRM with a data model you write yourself.
Attio matters to a directory like this one for a single reason: the objects are yours. Where every other CRM here makes an integration a mapping exercise against contacts and deals, Attio makes it a schema decision. The Clay pair is the clearest case, because the Object dropdown is populated from your own workspace and a custom object is as reachable as People or Companies.
It is also the only CRM in this directory that can run a whole enrichment inside its own automation builder. Send HTTP request calls the API, Parse JSON reads the answer, and the fields are written back with no middle layer, no second meter and no engineer. The ceiling is twenty seconds on the HTTP block, which a cached lookup fits comfortably and a live job does not.
The HubSpot route is the exception, and the page for it says so in its first line. Attio Import2 reads HubSpot and writes into Attio once, with a sample run and an undo, and there is no ongoing sync and no two-way option. Teams that import to evaluate Attio while still working in HubSpot end up maintaining two CRMs by accident.
Attio sits in the CRM layer, with 5 other tools in this directory. Counted across the whole directory it reaches 3 of the other 22 tools, and the integration matrix shows that row against every other.
Start with how the integration is built, because a native app and a webhook you maintain are the same feature with completely different owners.
Whichever of these you connect, the integration moves the record. We fill it.
Start for freeThe one CRM here whose own workflow can call the API and read the answer.
Not an integration. A one-time migration, and it will not update what it finds.
Write to objects you defined yourself, rather than to contacts and deals.
4 things that decide whether a Attio integration will work for you, in the order they usually bite.
Import2 runs once. There is no scheduled run, no ongoing sync and no two-way option, so treating it as an integration wastes a week. Records already in Attio are matched rather than merged: notes and tasks are added, and everything else on the record stays as it was. Undo removes the imported records and leaves behind the lists and attributes the migration created, so a second attempt starts in a workspace that is not clean.
Send HTTP request and Parse JSON together make Attio the only CRM here that completes an enrichment inside its own builder. The documented timeout is twenty seconds, which a cached lookup against a published 500ms P95 fits and a live job documented as possibly tens of minutes does not. Each extracted field declares an output type, and a mismatch fails the parse rather than returning empty. Attio also labels both blocks legacy while it reworks the workflow builder, so they work today and are not where the product is going.
Only an Attio workspace admin can authorise the Clay connection, which makes the first step a scheduling problem rather than a technical one. Import2 is admin-only as well, and available on Plus, Pro and Enterprise rather than on the free plan. Automations are a paid feature too, so the enrichment route needs a paid workspace even though the REST API does not. The free plan is enough to read the API and not enough to run anything on a schedule.
The published ceiling is 100 requests a second for reads and 25 a second for writes, and a single write can count as more than one request. Attio states that it may reduce the limit during incident response, and may permanently lower it for specific APIs that handle a lot of traffic. A limit that can move without you changing anything belongs in the retry logic rather than in a spreadsheet. A Clay lookup also returns at most ten matching records, which is enough to check whether somebody exists and not enough to reconcile a list.
Each one covers what syncs in which direction, the tier it needs, the limits that bite, when not to use it, and the questions people ask before they start.
The one CRM here whose own workflow can call the API and read the answer.
One way, in → 02Not an integration. A one-time migration, and it will not update what it finds.
One way, out → 03Write to objects you defined yourself, rather than to contacts and deals.
Two-way sync →The hub most sales stacks are actually built around.
See the 17 integrations →The widest integration surface, gated by edition.
See the 12 integrations →People and company data for every tool in this directory.
See the 22 integrations →Deep native CRM integrations, sold by tier.
See the 3 integrations →One integration that carries data, sequences and activities.
See the 7 integrations →An integration moves a record from one system to another. It does not make the record worth moving. Reverse Contact is the layer underneath: one API that resolves an email, a domain or a profile URL into a verified people and company record, with a confidence score instead of a guess.
Free trial, no card required, and 500 credits to start.
The questions that come up before anyone opens a setup guide.
Not in the sense most people mean. Attio Import2 reads HubSpot and writes into Attio once, as a migration, with a sample run to check the mapping and an undo afterwards. There is no ongoing sync, no scheduled run and no two-way option, so it is a way to leave HubSpot rather than a way to run both. It needs Attio Plus, Pro or Enterprise and a workspace admin to run it.
Yes, and it is the only CRM in this directory that can then read the reply and write it back without a middle layer. Send HTTP request has a documented twenty-second timeout, and Parse JSON extracts fields with a declared output type that fails rather than returning empty on a mismatch. Automations are a paid feature, and Attio currently labels both blocks legacy while it reworks the workflow builder.
The published ceiling is 100 requests a second for reads and 25 a second for writes, with a single write able to count as more than one request. Attio also states that it may occasionally reduce the limit as part of incident response, and may permanently lower it for specific APIs that handle a lot of traffic, so it is a ceiling that can move without anything changing on your side.
Clay Growth or above, an Attio workspace, and an Attio workspace admin to authorise the connection, which Clay documents as an admin-only action. The Object dropdown is populated from your own workspace, so custom objects work as well as People and Companies do. Two limits are worth knowing before you start: a lookup returns at most ten records, and an import stops at 50,000.
Tier requirements, sync behaviour and limits read from each vendor's own documentation in September, October and November 2026. Vendors change these without notice.
Reverse Contact appears in this directory as a data provider and publishes it. Every other tool is here on merit and nobody pays for placement.