Start with the tool you already run. Each one opens a hub listing every integration we have checked for it, with the objects that move, the direction they move in, the plan you need on each side, and the limit that will find you in week two.
POST/v2/enrich/personsAuthorization:Bearer $RC_API_KEY{"email":"j.okafor@northwind.io"}# the integration moved the record.# this is what fills it in.
fullNameJane Okafor
currentPositionVP Sales, Northwind
employeeCount220
confidence0.94
200 OKResponse time, P95500ms
68integrations checked
28need a plan upgrade
8worked workflows
5cost nothing extra
Start here
Pick the tool you already run
All 23 tools a B2B sales stack is built from, grouped by the layer they sit in rather than by category name. Ten have a full hub listing every integration we have checked for them. The other thirteen are documented here, and their guides live on the hub of the tool they connect to.
Every record was checked against the vendor's own knowledge base, developer documentation or pricing page in September and October 2026. Where a tier requirement is stated by the vendor, it is quoted. Where nothing is published, the record says so rather than guessing.
The data layer
Whichever of these you connect, the integration moves the record. We fill it.
More integrations point at HubSpot than at any other tool in this directory, which is the real reason teams pick it and the real reason they struggle to leave. The catch is that the automation that makes those integrations useful sits behind Professional, so the marketplace listing and the working integration are two different purchases.
Workflows, the thing most integrations trigger from, start at Professional
A private app you build gets 190 requests per 10 seconds; a marketplace app gets 110
Free CRM API access on every tier, including the free one
Almost every tool here ships a Salesforce connector, because enterprise buyers demand one. What varies wildly is whether your edition can use it: API access itself is an edition feature, and several integrations that look standard quietly require Enterprise.
API access is an edition feature, not a universal one
Group Edition is not supported by the HubSpot connector at all
Selective sync between HubSpot and Salesforce needs a Salesforce Enterprise licence
Attio matters to an integration directory for one reason: the objects are yours. If your motion runs on something that is not a contact and a deal, every other integration on this site is a mapping exercise and Attio's is a schema decision.
Objects and attributes defined through the API, not just in the interface
Ships an official MCP server alongside the REST API
Thinner native integration list than HubSpot or Salesforce
A pipeline a sales team will actually keep up to date
LayerCRM
Guides
Pipedrive is the CRM teams choose when they want the deal board and not the platform, which makes its integration story unusually legible: there is a well documented REST API, there are webhooks on every plan, and there is an automation builder that can fire one outward. What it cannot do is read the answer back, and that single gap shapes every enrichment design on it.
Webhooks on every plan, and a webhook action inside automations from Growth
A daily API budget in tokens rather than a flat call count
The automation webhook is global admin only, and cannot use the response
Zoho CRM is the only CRM in this directory that ships its own programming language. Deluge runs inside the product, calls external APIs with invokeUrl, and turns an enrichment build into a function rather than a pipeline. The catch is the edition: functions are fully available on Enterprise and above, and Standard and Professional reach them only through extensions.
Deluge functions call external APIs directly from a workflow rule
Functions in full only on Enterprise, CRM Plus, Ultimate and Zoho One
API limits are credits in a 24-hour window, not calls per minute
A sales team already living in the same boards as everyone else
LayerCRM
Guides
monday CRM is boards rather than objects, which makes it pleasant to change and awkward to enrich: there is no contact schema to map into, only columns somebody named. The API is GraphQL with a complexity budget, and the limit that decides an enrichment build is not the complexity but the daily call count, which is 1,000 until you reach Pro.
GraphQL API with a complexity budget and a separate daily call limit
1,000 API calls a day on Basic and Standard, 10,000 on Pro
Reverse Contact resolves an email, a domain, a name or a profile URL into a structured B2B record, and it does it behind an API rather than a marketplace listing. That is a deliberate shape: it is the layer you put underneath HubSpot, Clay or your own product, not another app competing for a seat in the stack.
One API for search, enrichment and live on-demand resolution
Returns null with a confidence score rather than a best guess
Ships an MCP server with 37 tools, so an agent can call it directly
ZoomInfo's CRM integrations are genuinely deep and genuinely tiered. Inbound Enrich fires the moment a record is created, bulk enrichment keeps the back catalogue current, and which of those you get depends on which edition you bought.
Native HubSpot sync starts at ZoomInfo Advanced
Two-way sync and intent fields are an Elite feature
Configurable field mapping, so you control which fields it overwrites
Apollo is unusual here because its CRM integration moves three different things at once: contact and account records, sequence activity, and lists in the other direction. That breadth is the whole argument for it, and the reason its sync settings deserve more care than most.
Bi-directional push and pull of contacts, accounts and deals
Sequence activity logs to the CRM timeline as calls, emails, meetings, tasks and notes
Clay is the integration nobody lists as an integration. It reads records out of your CRM, runs them through however many providers you point at, and writes the result back, which makes it the fastest way to connect a data source that has no marketplace app. Reverse Contact is one of those.
Lookup actions read from the CRM without changing anything
Create and Update actions write back, visible to every CRM user immediately
Any HTTP API becomes a column, which is how tools without a connector get one
n8n matters here because it removes the per-task meter and the data-residency argument in one move. If an integration in this directory has no native app and your security team will not accept a third-party cloud in the path, this is usually the answer.
Self-hostable, so records never leave infrastructure you control
An HTTP Request node covers any API without a dedicated node
Native HubSpot and Salesforce nodes for the common objects
When a pair in this directory has no native integration, Zapier usually has both sides. You pay for that in tasks rather than seats, which is fine at low volume and stops being fine the moment an enrichment loop starts running per record.
HubSpot alone reaches 8,000+ apps through Zapier
The Zapier app has 173,000 installs on the HubSpot marketplace
Billed per task, so a per-record loop is the expensive shape
Make earns its place in this directory on one specific job: custom field handling that native integrations skip. It comes up most often as the fallback when a native connector covers the standard objects and stops at the fields you actually filter on.
Full custom-field support where several native connectors have none
Visual branching, so one scenario can fan out per record type
Billed in operations, which counts every module run, not every record
Sales Navigator has two quite different integrations with a CRM and they are routinely confused. The embedded profile card is one thing. CRM Sync, which writes activity back and creates records, is another, and it only exists on Advanced Plus.
CRM Sync requires Sales Navigator Advanced Plus, not Advanced
Syncs Contacts to Leads, Companies to Accounts, Deals to Opportunities
Field mappings cannot be customised: everything syncs or nothing does
Outreach treats the CRM sync as part of the product rather than a bolt-on, with per-object mapping and polling rules you configure rather than accept. That is the difference you are paying for against a cheaper sender.
Per-object, per-field mapping in both directions
Sequence activity writes back as CRM activities
Governance controls over which side wins a conflict
Smartlead's integration story is webhooks rather than connectors. Every reply, bounce and unsubscribe fires an event you can route anywhere, which is what you actually want when the CRM needs to know within seconds rather than at the next sync.
Event webhooks for sends, opens, replies, bounces and unsubscribes
Campaign creation is a real API operation, not a read-only view
Native in Clay, so a list can go from table to campaign in one step
lemlist runs email and LinkedIn touches in the same sequence, which is why its CRM integration has to carry more activity types than a pure sender. Worth knowing: lemlist's own head of product is on record as a Reverse Contact customer, so the data layer under it is the same one in this directory.
Enterprise outbound with a rate limit the whole team shares
LayerEngagement
Guides
Salesloft is bought the way Outreach is bought, by a revenue operations team with a procurement process, and it integrates the same way: an API you are a customer of rather than a platform you build on. The detail that catches enrichment builds is that the rate limit is a budget for the whole team, so the integration you add competes with every integration already there.
A cost-based rate limit of 600 a minute, shared across the whole team
Webhooks on 37 event types, including person and cadence membership
API keys are for customers only; partners must ship an OAuth app
High-volume cold email with the API limits to match
LayerEngagement
Guides
Instantly is the sequencer in this directory with the most generous API limits by a wide margin, 6,000 requests a minute against Salesloft's 600 cost, which makes a bulk enrichment pass a normal thing to do rather than something to schedule around. It also documents a data-provider registry that can run on a key you bring, though the public documentation names no providers.
100 requests a second and 6,000 a minute, across the whole workspace
Bearer-token API v2, with scoped keys
A documented provider registry that can bill a workspace key at zero
Slack is the integration that decides whether the rest of the stack gets used. A pipeline that enriches, scores and routes perfectly still fails if a rep finds out about it tomorrow. The limits here are about routing, not about the connection.
A HubSpot workflow can send to a static channel, not to the channel on the record
Deal and company records can each be associated with a channel
Creating a channel from a workflow needs HubSpot Professional or Enterprise
The reason this pair comes up constantly is handover. A deal closes in the CRM and something has to happen in the place delivery actually lives. ClickUp is the one that put both halves of that in place: its actions run inside HubSpot workflows, and HubSpot actions run inside ClickUp automations.
ClickUp actions available inside HubSpot workflows
HubSpot actions available inside ClickUp automations
Gong is a one-directional integration by design: it reads the calendar and the CRM, and it writes intelligence back. The value is entirely in the write-back, so the field mapping is the whole configuration.
Call, email and meeting data attached to CRM records
Deal-level signals written to CRM properties
Read access to the CRM is required before anything writes back
Not a sales tool, and the most consequential integration in most stacks anyway. Gmail and Calendar are the top two installs on HubSpot's marketplace by a wide margin, because they are the ones that decide whether activity data exists at all.
Gmail is the most-installed app on the HubSpot marketplace, at 532,000 installs
Google Calendar follows at 289,000, ahead of Outlook at 247,000
Sheets remains the most common handover format between tools that will not talk
A stack is a pipeline, so the useful question about any tool is what comes before it and what comes after, not which quadrant it appears in.
Most tool directories sort by category, which tells you what a product calls itself and nothing about whether it belongs in your stack. This one sorts by layer, because the only question that matters about a tool is what it receives and what it hands on.
A typical B2B motion runs through five of the seven layers here. Prospecting produces a list, data fills it in, the CRM stores it, engagement acts on it, and collaboration tells a human when to intervene. Orchestration is the layer that connects the other five when no native integration exists, which is why Clay, n8n, Zapier and Make show up in nearly every stack on this site despite doing none of the actual work.
Revenue intelligence is the one that sits outside the flow. Gong reads the CRM and the calendar and writes back, so it is not a step in the pipeline, it is an observer of it. That is worth knowing before you try to place it in a diagram.
The seven layers are also how this directory is queryable rather than only readable: every record carries the layers it spans, and the MCP server answers on them, so an agent can ask what crosses the edge between two layers without reading a page at all.
The system of record. Everything else in the stack is either filling it, reading it or reacting to it, which is why the CRM choice constrains every other integration decision you make afterwards.
Sequencing, sending and reply handling. The layer with the most vendors and the least differentiation, where integration quality is usually the deciding factor.
The glue. Clay, n8n, Zapier and Make exist because the native integration you need often does not, and 25 of the 61 pairs in this directory depend on one of them.
Slack, ClickUp, Google Workspace. Where a human finds out something happened. Cheap to connect and the single most common place a working stack fails quietly.
Call and meeting data written back onto the record. An observer rather than a step, and the only layer here where the integration is entirely one-directional by design.
Head to head
The data layer, compared on how it connects
Layer two of a sales stack is the one you bolt onto the CRM to fill the records it already holds. Four contenders, compared on the things that decide an integration rather than on database size.
This is the row of the stack Reverse Contact competes in, and it publishes this directory, so read the table with that in mind. It is built from the same records as every other page here, including the rows where Reverse Contact is the weaker answer.
Enrichment on record creation, at the enterprise end
Data and sending behind one connector
Logic that changes monthly
Read from each vendor's own documentation and pricing pages in September and October 2026. "Not published" means exactly that: the vendor does not state the behaviour, and we are not going to infer it. Vendors move these quietly, so confirm against your own contract.
Buyer's guide
How to read an integration listing
Six things that decide whether an integration will work for you, in the order they usually bite. Every figure below is counted from the records in this directory.
01
"Integrates with" is not a capability
Every tool in this directory claims to integrate with every other one, and the claim is usually true in the sense that an integration exists. What varies is whether it moves records or only events, whether it goes both ways, and whether the useful half is on the edition you own.
LinkedIn Sales Navigator is the clearest example. The embedded profile card and CRM Sync are both "the HubSpot integration". One shows a profile. The other writes activity back, creates contacts and matches records, and it costs an edition upgrade.
02
The plan requirement is the real price
28 of the 68 integrations here need a higher plan or a separate licence on one side before anything works, and only 5 cost nothing extra on either side. The integration itself is nearly always free. Getting to the tier that carries it is the bill.
Price the tier, not the connector. If the upgrade across your seat count costs more than the problem you are solving, the correct answer is a different route rather than a cheaper plan, because the cheaper plan will not carry the feature.
03
Ask what happens to the history
A meaningful number of integrations sync forward only. LinkedIn CRM Sync states it plainly: only new and updated data syncs after the sync is turned on, and getting history across means uninstalling and reinstalling the app.
That is a decision about your switch-on date rather than a bug, and it is much easier to accept before the install than in the meeting where the first report comes back empty.
04
Whoever writes last wins, so decide who writes
20 of these sync both ways, which sounds like a feature and is really a governance question. Two systems writing the same field is two systems disagreeing on a schedule. Every unpleasant integration story in this category starts there.
The pattern that survives is boring: enrichment writes to its own properties, humans write to theirs, and a workflow promotes a value only where the human field is empty. It takes an extra afternoon and removes an argument that otherwise arrives in month three.
05
Native does not mean maintained by someone else
28 of these are native and 12 are something you build. The native ones are easier to start and no easier to own: field mapping, tier changes and sync conflicts are still yours. What native buys you is that the connection itself will not break when an API version moves.
Where no native option exists, the honest comparison is not "app versus no app". It is a day of work against an ongoing tier cost, and the day of work often wins.
06
MCP is starting to replace the small integrations
Two records here ship an MCP server, and that number will not stay at two. An agent that can call a data API and a CRM directly does not need the integration between them to exist, which removes the build for a class of jobs that used to justify one.
It does not replace a sync. A scheduled two-way pipeline is still a pipeline. What it replaces is the one-off research task somebody was about to automate badly. There is a worked example.
Workflows
Integrations assembled into something that runs
An integration on its own is a connection. A workflow is a decision about what happens to a record from the moment it appears to the moment a human acts on it. Eight worked sales stacks, one page each.
Each row opens the HubSpot hub filtered to the integrations that carry what that job needs, because HubSpot is where most of these questions start.
Nobody sets out to buy an integration. They set out to stop retyping something, and the integration is what that turns into. Reading by the job rather than by the logo usually shortens the shortlist to two.
This is a shortlist, not a census. We started from the tools that come up when a B2B team assembles a sales stack, then cut it to pairs where at least one side publishes documentation specific enough to check.
Every tier requirement, sync direction and limit here was read from the vendor's own knowledge base, developer documentation or pricing page between September and October 2026, not from a comparison article. Each guide names its sources at the bottom and the date it was read. Vendors move these quietly, and Reverse Contact itself moved three of them in three weeks, so treat every figure as the start of your own check. The full methodology sets out the five things this directory refuses to publish and the routes for checking any of it without asking us.
Where a vendor publishes no public figure, we say so. Reverse Contact publishes what each call costs in credits, 2 for a reverse lookup from an email and nothing for a freshness check, and publishes no dollar rate behind them, so every page here states the credit and leaves the conversion to the pack or plan you are on. Dividing a pack price by a credit count would be our arithmetic wearing their authority. ZoomInfo, Outreach, Gong and Sales Navigator Advanced Plus are quoted rather than listed, and the pages say that instead of estimating.
Reverse Contact publishes this, and is in it
This directory is published by Reverse Contact, which also appears in it as a data provider. That is a conflict worth stating rather than hiding. Reverse Contact's own pages are held to the same standard as everyone else's, including the part where it has no marketplace app on any CRM and we say so on every page where it comes up. It is also the weaker answer in several rows of the comparison above, and those rows stayed in. No other vendor pays for placement, and outbound links carry nofollow.
No scores, no stars, no leaderboard
We publish no rating out of five and no aggregate score, because an integration is not better or worse in the abstract. It is right or wrong for a specific tier, a specific object and a specific direction, and those are the fields we publish instead.
Limits are quoted, never softened
Where a vendor documents a limit, it appears here in the vendor's own terms, including the awkward ones. HubSpot's Slack workflow action cannot route to the channel on the record. LinkedIn's CRM Sync will not backfill. ClickUp's native path thins out at custom fields. An integration directory that only lists capabilities is a marketing page with a filter on it.
Connecting the tools is half the job
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, under whichever tools you just connected.
The ones that come up in almost every conversation about assembling a stack.
What counts as an integration here?
Any supported route by which two tools exchange records or events: a native connector built by one of the vendors, a certified marketplace app, an iPaaS scenario in Zapier, Make or n8n, something you build on the two APIs, or an MCP server an agent can drive. We list the route rather than the marketing claim, which is why several entries here say "you build it" on a pair that both vendors describe as integrated.
Which integration should I set up first?
Whichever one stops a person retyping something today. In practice that is almost always email and calendar into the CRM, because without activity data every other integration is reporting on a partial record. After that it splits: if inbound is your motion, enrichment on record creation comes next; if outbound is, it is the sequencing tool's CRM sync. Slack comes last and matters more than its price suggests.
Why do so many of these need a plan upgrade?
Because integrations are a retention feature and vendors know it. A connected tool is a tool nobody removes, so the integration is offered widely and the useful half of it is placed one tier up. 25 of the 61 here need a higher plan or a separate licence on one side. That is not a criticism of any single vendor, it is the shape of the category, and it is why we put the tier requirement on the card rather than three screens into a setup guide.
Native app or build it myself?
Native when it exists, fits your tier and moves the objects you need. Build it when any of those three is false. The thing teams get wrong is assuming a native app removes ownership: you still own the field mapping, the conflict rules and the decision about what may overwrite a human. What native actually buys is that the connection will not break when an API version changes, which is worth real money and is not the same as being maintenance-free.
Is Zapier good enough?
For a trigger that fires a few hundred times a month, yes, and it will be running long after a custom script has been forgotten. It stops being good enough when the shape is per-record rather than per-event, because the task meter then scales with your data rather than with your activity. An enrichment loop across 20,000 CRM records is the standard example of a job that looks like a Zap and should not be one.
Where does Reverse Contact fit in a stack?
Underneath it. Reverse Contact is not another tile in the CRM, it is the API that fills the record the integration just moved: an email, a domain or a profile URL in, a verified people and company profile out, with a confidence score and a null rather than a best guess when it is not sure. There is no marketplace app on any CRM, which is a design decision rather than a gap, and the four routes into HubSpot are worked through on its own page.
How often is this checked?
Every record carries the date it was verified, and those run from September to October 2026 rather than one date across the board. Tier requirements are the figure most likely to move, because they move whenever a vendor repackages, and they rarely get announced. If you find one that has changed, tell us and we will correct it rather than leave it standing.
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.