Salesforce integration tools compared: the best Salesforce data integration and ETL software
Twelve ways to connect Salesforce to the rest of your stack, lined up by who owns each tool, how it bills, and who it is wrong for. Two things decide this purchase more than any feature list: the Salesforce API allocations your edition actually gets, and the five free Integration user licenses most orgs never claim. Both are tabled below, read from Salesforce documentation on 19 August 2026.
Field mapping auto-plugged · tap a port to rewire
Plug a source port into
Transform on this cable
JSON in
JSON out
5 sample records ready
Salesforce documentation read 19 August 2026 · Last updated August 2026
Which Salesforce integration tool should you use?
Match the tool to the job, not to the logo. If the goal is Salesforce data in a warehouse for reporting, that is ELT and Fivetran, Airbyte or Adapters do it. If the goal is keeping Salesforce and an ERP in agreement, that is iPaaS and Boomi, Workato, Jitterbit or Celigo do it. If the goal is pushing modeled scores from the warehouse back onto Salesforce records, that is reverse ETL and Hightouch or Adapters do it. MuleSoft is the deepest option and belongs to organizations with a dedicated integration practice and a budget to match. And for a meaningful number of teams the right answer is no vendor at all, because Flow with External Services, Change Data Capture and the Bulk API are already included in the edition you pay for.
The correction worth making early: most published comparisons of this category rank tools on connector counts and never mention the two constraints that actually decide the project. Your edition sets a hard daily API allocation shared across every integration you run, and only 25 requests lasting 20 seconds or more can execute at once. A tool that polls aggressively will hit those ceilings no matter how good its interface is. For the wider category outside Salesforce, start with best data integration tools.
Salesforce integration tools compared
Billing units rather than price tags, because most of this category is quote-based or consumption-metered and any number we print goes stale. The one flat published figure below is ours. Verify current pricing with each vendor before you sign anything.
| Tool | Owner | Approach | Best for | Pricing model | Watch out for |
|---|---|---|---|---|---|
| MuleSoft | Salesforce | API-led integration, full lifecycle API management | Enterprise architecture teams with a dedicated integration practice | Quote-based annual contract, billed on flows and messages | The heaviest option to learn and staff. Rarely justified for a handful of syncs |
| Data Cloud | Salesforce | Zero copy federation and sharing with cloud warehouses | Orgs already on Data Cloud that want warehouse data visible in Salesforce | Consumption credits on top of a Data Cloud subscription | Federation does not write values onto standard Salesforce object fields |
| Native platform | Salesforce | Flow, External Services, Platform Events, Change Data Capture, Bulk API | Teams with admin or developer capacity and a small number of endpoints | Included with your edition. No separate integration bill | You own the retries, the error log and the person who fixes it at 2 AM |
| Workato | Independent | Enterprise automation with a large recipe library | Mid-market and enterprise ops teams automating across many applications | Annual contract priced on workspaces and recipes | Entry commitments start well above typical departmental budgets |
| Boomi | Independent | Established low-code iPaaS with strong ERP coverage | Companies integrating Salesforce with NetSuite, SAP or on-premises systems | Per connection subscription, plus a pay-as-you-go option billed per message | Connection-based pricing climbs quickly as the endpoint count grows |
| Jitterbit | Independent | iPaaS with prebuilt Salesforce process templates | Salesforce to ERP projects that want a starting template rather than a canvas | Tiered annual subscription | Template fit is excellent or irrelevant, with little middle ground |
| Celigo | Independent | Prebuilt integration apps, strongest around NetSuite | Salesforce and NetSuite quote-to-cash in particular | Subscription by tier and endpoint count | Value is concentrated in the prebuilt apps, less so in custom work |
| Zapier | Independent | Trigger and action automation across thousands of apps | Lightweight record-level automation owned by a business user | Tiered by tasks executed per month | Task metering makes high-volume Salesforce syncs expensive and slow |
| Fivetran | Independent | Managed ELT into a warehouse with schema drift handling | Analytics teams loading Salesforce objects into Snowflake or BigQuery | Monthly active rows for connections, model runs for transformations | One direction only. It fills the warehouse, it does not write back |
| Airbyte | Independent | Open source ELT, self-hosted or cloud | Engineering teams that want control over where the pipeline runs | Free self-hosted, or cloud billed by credits tied to rows and volume | Self-hosting is free in license and not free in engineering time |
| Hightouch | Independent | Reverse ETL, warehouse back into Salesforce | Teams pushing modeled scores and segments onto Salesforce records | Tiered by destinations and synced records | The opposite direction to Fivetran. Most teams eventually want both |
| Adapters | Independent | Scheduled two-way syncs with visual field mapping and no-code transforms | Teams that need a handful of dependable Salesforce syncs on a predictable bill | Flat monthly, $49 to $399. No row meter, no task counter, no per-connector fee | Not the right tool for forty obscure connectors or distributed Spark jobs |
The column a table cannot show you is connector depth. Two vendors will both list Salesforce and only one will replicate the custom object with the formula field your forecast actually depends on. During the trial, sync the exact objects you need, including the custom ones, and read what landed rather than trusting the logo on the connector page. If you are evaluating against a specific vendor, we keep honest write-ups of MuleSoft, Workato, Boomi and Zapier.
What Salesforce already gives you before you buy anything
Six integration mechanisms ship with the platform. Knowing which one fits changes the shortlist, because a vendor whose connector polls the REST API every five minutes is spending your daily allocation on a job that Change Data Capture would do for free.
| Mechanism | What it does | Reach for it when | What it costs you |
|---|---|---|---|
| REST API | Record-level reads and writes | Anything conversational: create one Opportunity, look up one Account | Counts against the daily allocation, one call per request |
| Bulk API 2.0 | High-volume asynchronous load and extract | Anything above a few thousand records, in either direction | Far cheaper per record against the allocation than REST |
| Change Data Capture | Streamed change events for created, updated, deleted and undeleted records | You want to react to changes instead of polling for them | Event delivery allocations rather than API calls |
| Platform Events | Custom publish and subscribe messages you define | Event-driven architecture where you control the payload shape | Event delivery and publishing allocations by edition |
| External Services and Flow | Call an external REST API from a Flow with no code | An admin can own the automation and the endpoint is well behaved | Included. Callout limits apply per transaction |
| Salesforce Connect | External objects surfaced live, without copying data in | You need to see external data, not own it | Separate add-on licensing |
The honest limit of the native route is not capability, it is ownership. Flow can call an external service and Bulk API can move a hundred million records, but neither ships with retries you did not write, a per-record error log somebody can read, or an alert that reaches a human. That gap is what you are buying when you buy a tool, and it is worth being explicit about it rather than discovering it during the first failed load. The same trade is covered generally in build versus buy for integrations.
Salesforce API limits every integration has to live inside
These are shared across every integration in the org, not allocated per tool. Read from Salesforce developer documentation on 19 August 2026. If you take one table from this page, take this one.
| Limit | Value | Why it matters |
|---|---|---|
| Daily API calls, Enterprise and Professional | 100,000 + 1,000 per Salesforce license | Shared across REST, SOAP, Bulk and Connect. Purchased add-ons raise it |
| Daily API calls, Unlimited and Performance | 100,000 + 5,000 per Salesforce license | The per-license rate is what makes large orgs comfortable, not the base |
| Daily API calls, Developer Edition | 15,000 | Enough to build against, not enough to load test against |
| Daily API calls, full sandbox | 5,000,000 | A sandbox will never reproduce the production limit you are about to hit |
| Concurrent long-running requests | 25 in production, 5 in Developer and trial orgs | Applies to requests lasting 20 seconds or longer. The quietest cause of integration failures |
| API request timeout | 10 minutes | REST and SOAP calls. Query calls follow SOQL limits instead |
| Request URI and headers | 16,384 bytes combined | Long SOQL sent as a GET query string can breach this before you expect it |
| Bulk API 2.0 daily ingest | 150,000,000 records per 24 hours | The reason bulk loads should never be a loop of REST calls |
| Bulk API 2.0 job payload | 150 MB of base64-encoded CSV | Split larger files before submitting, not after the job fails |
| Bulk API 2.0 record and field size | 400,000 characters per record, 131,072 per field | Long text and rich text fields are where this bites |
| Bulk API 1.0 batch | 10,000 records or 10 MB per batch | Still in use in older integrations. Bulk API 2.0 handles batching for you |
The daily allocation is a soft limit in practice, so a brief overrun usually produces a warning rather than an outage. Sustained overrun does not: calls start returning REQUEST_LIMIT_EXCEEDED and every integration in the org fails together, including the ones nobody was changing that week. That shared blast radius is the argument for one integration user per integration, because it is the only way to tell which system spent the allocation.
The five free Salesforce Integration user licenses most orgs never claim
Salesforce introduced an API-only user license specifically for system-to-system integrations, and Enterprise, Unlimited and Performance orgs are provisioned five of them at no additional cost. Teams routinely burn a full paid license on an integration instead, or worse, run the integration as a named admin whose deactivation later takes the pipeline down with them.
| Detail | What applies |
|---|---|
| Free allocation, Enterprise, Unlimited, Performance | Five Integration user licenses at no additional cost |
| Free allocation, Developer Edition | One Integration user license |
| User interface access | None. The user cannot sign in to Salesforce or view data in the UI, only call the APIs |
| Profile | Salesforce API Only System Integrations |
| Permission set license | Salesforce API Integration, assigned before any permission set |
| Object access | Granted through permission sets built on the Salesforce API Integration license |
| Common gotcha | Permission sets on this license cannot contain Visualforce pages or Lightning Experience user permissions |
| Why it matters | One integration user per integration makes an audit trail readable and a failure traceable to one system |
The order of operations trips people up: assign the Salesforce API Integration permission set license to the user first, then build permission sets against that license, then assign them. Doing it the other way round produces permission sets that will not attach. Full walkthrough in the Salesforce Integration user license guide.
Six numbers worth knowing before the first demo
5 free
Integration user licenses on Enterprise and above
Enterprise, Unlimited and Performance orgs are provisioned five Salesforce Integration user licenses at no extra cost. Developer orgs get one. Most teams paying for a full license to run an integration do not know these exist.
100,000
Base daily API calls on Enterprise
Plus 1,000 per Salesforce license, and 5,000 per license on Unlimited and Performance. The pool is shared across REST, SOAP, Bulk and Connect APIs. Source: Salesforce developer documentation.
25
Concurrent requests lasting 20 seconds or more
Production and sandbox orgs alike. Developer and trial orgs get 5. This ceiling causes more mysterious integration failures than the daily allocation does, because it is invisible until a slow query stacks up.
150M
Bulk API 2.0 records per 24 hours
The daily ingest allocation. A single job payload is capped at 150 MB of base64-encoded CSV, with 400,000 characters per record and 131,072 per field.
10 min
REST and SOAP API request timeout
Anything that needs longer belongs in Bulk API 2.0, which is asynchronous by design. Retrying a timed-out synchronous call is how teams quietly burn their daily allocation.
$49/mo
Adapters entry price, flat
No task meter, no monthly active rows, no per-connector charge. The bill is the same whether the sync moves 400 Salesforce records or 400,000.
How to set up a Salesforce integration that survives contact with production
Step 01 / User
Create a dedicated integration user before anything else
The single most common mistake in a Salesforce integration is authenticating as a real person. When that admin leaves, or their permissions change, or they get deactivated during an offboarding, the integration dies and nobody connects the two events for a day. Create one Integration user per integration, on the Salesforce API Only System Integrations profile, with the Salesforce API Integration permission set license assigned before any permission set. Enterprise, Unlimited and Performance orgs get five of these at no cost, so on most orgs this step is free.
Step 02 / API
Pick the API by volume, not by familiarity
REST is comfortable and it is the wrong choice above a few thousand records, because every record costs a call against a shared daily pool. Bulk API 2.0 is asynchronous, handles batching itself, and ingests up to 150,000,000 records per 24 hours. If the requirement is reacting to change rather than moving volume, Change Data Capture and Platform Events replace polling entirely, and polling is what quietly consumes an allocation overnight.
Step 03 / Map
Decide the field mapping and the conflict rule in the same sitting
Field mapping is the easy half. The half that causes arguments six weeks later is what happens when both systems changed the same record: last write wins, Salesforce always wins, or the other system always wins. Pick per object rather than globally, because the correct answer for Account billing address is rarely the correct answer for Opportunity stage. Write it down where the next person will find it.
Step 04 / Match
Choose an external ID and use upsert
Insert-only integrations create duplicate Accounts until somebody notices, and duplicate Accounts are expensive to unwind once activity has attached to both. Mark a field as an External ID in Salesforce, populate it from the source system primary key, and upsert against it. This is the difference between an integration that survives a replay and one that doubles your data when a job runs twice.
Step 05 / Watch
Alert on the sync that stopped, not just on the error that fired
Two failure modes matter and only one of them is loud. A rejected record throws an error you can catch. A sync that has silently not run since Tuesday throws nothing at all, and the dashboard keeps rendering last week numbers as if they were current. Monitor freshness and record counts, watch the concurrent long-running request ceiling of 25, and route the alert to a person who is on duty rather than to a channel nobody reads.
What teams actually connect Salesforce to
Salesforce into a warehouse for reporting
The most common request by a distance. Salesforce reports cannot join to product usage, billing or support data, so Account, Opportunity, Lead and the custom objects get landed in Snowflake, BigQuery or Postgres and modeled there.
Salesforce and your ERP in agreement
Closed Won creating a sales order in NetSuite, and the resulting invoice status coming back onto the Opportunity. Quote-to-cash is where a broken integration becomes a finance problem rather than a reporting one.
Marketing automation both ways
Leads flowing from HubSpot or Marketo into Salesforce, and lifecycle stage flowing back so marketing stops emailing closed customers. Bidirectional by nature, which is why the conflict rule matters here more than anywhere else.
Billing and payments onto the account record
Stripe subscription status, invoice history and payment failures visible on the Account, so a rep is not asking for a renewal from a customer whose card declined last week.
Support tickets alongside the pipeline
Zendesk or ServiceNow ticket volume and severity on the Account, so churn risk is visible before the renewal call rather than after it.
Modeled scores back into Salesforce
Health scores, propensity models and usage tiers computed in the warehouse and written onto Salesforce records. That is the reverse direction, with its own failure modes around API limits and upsert matching.
The first one has a dedicated walkthrough in syncing Salesforce to Snowflake, and the last one, pushing modeled results back onto Salesforce records, is covered on reverse ETL tools. For the marketing case specifically, see Salesforce to HubSpot, and if HubSpot rather than Salesforce is the system of record on your side of the sync, the equivalent allocations and native options are tabled on HubSpot integration tools. Teams attaching purchase history to the Salesforce record usually need the commerce side mapped first, which is on Shopify integration tools.
When another Salesforce integration tool is the right answer
We would rather you buy the correct thing than churn in month three. Five situations where Adapters is not what you want.
The native platform already covers it
If the requirement is one Flow calling one well-behaved REST endpoint, and an admin on your team can own it, that is included in the edition you already pay for. Buying a platform for that is spending money on a solved problem.
You need full API lifecycle management
Designing, versioning, publishing and governing APIs for other teams to consume is MuleSoft territory, and it is genuinely good at it. We move data between systems. We do not manage an API estate.
You need dozens of connectors
If the list runs to forty SaaS sources including several obscure ones, buy catalog breadth. Workato, Boomi, Fivetran and Airbyte have spent years on that surface area and we have not.
The logic belongs in Apex
Branching business rules that have to run inside the Salesforce transaction, before a record commits, belong in Apex and Flow. An external sync runs after the fact by definition, and no scheduling frequency changes that.
Your volume genuinely is enormous
Flat pricing is a bet that most teams move a predictable amount of data. If you are moving hundreds of millions of Salesforce records a month, negotiate a volume contract with a metered vendor and you will likely beat our list price. Our advantage is the middle, where a metered bill is unpredictable and a hand-built integration is a permanent staffing cost nobody put on the budget.
Four questions to ask on every Salesforce integration demo
Question 01
Which API does it use, and how many calls per sync?
A vendor that cannot answer this is spending an allocation it does not track. You want Bulk API 2.0 for volume and Change Data Capture instead of polling wherever possible.
Question 02
Does it support custom objects and custom fields?
Every vendor lists Salesforce. Not every vendor replicates the custom object with the formula field your forecast runs on. Test the exact objects during the trial.
Question 03
How does it match records on update?
You want upsert against a field marked External ID. Insert-only behavior creates duplicate Accounts, and duplicates are expensive to unwind once activity has attached to both.
Question 04
What happens when 40 records fail validation?
Salesforce validation rules reject records the source system considered fine. You want the per-record error log with the values attached, the retry behavior, and the alert path.
Where Salesforce data usually goes next
Most Salesforce integration projects end up as one of these specific pairs. Each page covers the object mapping, the type conversions and the failure modes for that route.
Questions buyers ask about Salesforce integration tools
- What are Salesforce integration tools?
- Salesforce integration tools are the platforms that move records between Salesforce and the rest of your stack, either by calling the Salesforce APIs on a schedule or by reacting to events. They fall into three groups: Salesforce-owned software such as MuleSoft and Data Cloud, general iPaaS vendors such as Workato and Boomi, and warehouse-focused ELT vendors such as Fivetran and Adapters.
- What is Salesforce integration?
- Salesforce integration is the practice of keeping Salesforce and another system in agreement without anyone retyping data. In practice it means authenticating against the Salesforce API, mapping fields in both directions, deciding which system wins on a conflict, and running that on a schedule or trigger with retries and error logs.
- What are the best Salesforce integration tools?
- There is no single best one, only a best fit. MuleSoft is the deepest option and the most expensive. Workato and Boomi lead the general iPaaS field. Zapier wins on breadth for lightweight automations. Fivetran and Airbyte lead warehouse loading. Adapters fits teams who want a handful of dependable Salesforce syncs on a flat monthly bill.
- How much does Salesforce integration cost?
- The software is only part of it. Salesforce itself charges nothing extra for API access on Enterprise, Unlimited and Performance editions, and gives those orgs five Integration user licenses free. The integration platform is the variable: entry iPaaS plans run in the hundreds of dollars a month, MuleSoft enterprise contracts run into six figures a year, and Adapters is flat at $49 to $399 a month.
- Does Salesforce have its own integration tool?
- Several. MuleSoft Anypoint Platform is the flagship, acquired in 2018. Data Cloud offers zero copy federation and sharing with warehouses. Inside the platform itself, Flow with External Services, Platform Events, Change Data Capture and the Bulk API all handle integration work with no third-party vendor at all.
- What is a Salesforce integration user license?
- It is an API-only user license for system-to-system integrations. A user holding one cannot sign in to the Salesforce user interface or view data there, only call the APIs. Enterprise, Unlimited and Performance orgs are provisioned five of them at no extra cost, and Developer orgs get one.
- What are the Salesforce API limits for integrations?
- Enterprise and Professional orgs get 100,000 calls per 24 hours plus 1,000 per Salesforce license. Unlimited and Performance get 100,000 plus 5,000 per license. Developer Edition gets 15,000 and a full sandbox gets 5,000,000. Separately, only 25 requests lasting 20 seconds or longer can run concurrently in production, and 5 in a Developer org.
- How do I integrate Salesforce with an external system?
- Create a dedicated integration user, authorize a connected app with OAuth, pick the API that matches the volume, then map the fields and choose a conflict rule. Use REST for record-level work, Bulk API 2.0 for anything above a few thousand rows, and Change Data Capture or Platform Events when you need to react rather than poll.
- What is ETL in Salesforce?
- ETL in Salesforce usually means extracting objects such as Account, Opportunity and custom objects into a warehouse for reporting, because Salesforce reports cannot join to the rest of your data. Most teams run ELT instead: land the raw objects in Snowflake, BigQuery or Postgres, then model them there with SQL.
- How do I integrate Salesforce with Snowflake?
- Two routes. A connector reads the Salesforce objects through the Bulk API and loads them into Snowflake on a schedule, which works with any edition. Or Data Cloud zero copy shares data with Snowflake without moving it. The connector route is more common because it does not require Data Cloud.
For the categories around this one, see iPaaS software when the job is keeping business applications in agreement, and ETL software for how our own pipeline works end to end.
Connect Salesforce to your stack on a flat monthly price
Map the fields once, pick a schedule, and let it run with retries, alerts and per-record logs. From $49 a month, with no task meter and no row counter to forecast.
No credit card required.