Quick answer: what Workato costs in 2026
Workato publishes no prices. Every deal is quoted through sales. The bill combines a platform edition fee, across four editions from Standard to Workato ONE, with a usage fee measured mostly in business actions. Extra on-premise agents (OPAs) and concurrency can be bought as additional capabilities. The per-unit rate appears only inside a quote.
Everything here was checked against workato.com/pricing and the pricing pages on docs.workato.com on September 14, 2026. No third-party price estimates appear anywhere below, because Workato does not confirm them. One caveat before you read further. Customers who joined Workato before February 2024 are on a separate, older pricing model. A renewal conversation with a long-standing account can look nothing like the model described here.
Workato editions at a glance
| Edition | Workbots | OPAs | Concurrency | Adds |
|---|---|---|---|---|
| Standard | 1 Standard Workbot | None listed | 1 | Core integration |
| Business MCP | 10 Enterprise | 3 | 1 | MCP servers, API tools |
| Enterprise MCP | Unlimited | 10 | 15 | Pipelines, AI, Insights |
| Workato ONE | Unlimited | 30 | 30 | Agent Studio, GO |
Those on-premise agent and concurrency counts are totals for the edition, not amounts added on top of the previous tier. An Enterprise MCP contract includes 10 on-premise agents in all, not 10 on top of the 3 in Business MCP. Features work the other way. Each higher edition includes everything in the one below it, which is why the Adds column lists only what is new.
How Workato's pricing model works
For direct customers on the current model, the documentation describes two components. One is a platform edition fee, which determines access to capabilities. The other is a usage fee that scales with usage volume, so what you pay tracks what you actually run. On top of an edition, customers can also buy additional capabilities such as extra on-premise agents and extra concurrency.
"Workato offers a usage-based pricing model for direct customers. This model includes: A platform edition fee, which determines access to platform capabilities. A usage fee, which scales based on your usage volume. You only pay for what you use."
Source: docs.workato.com
Two qualifiers matter for budgeting. Pricing and plan structures may vary based on contract terms and when the customer joined. That is why two companies on the same edition can sit on different structures. And the usage-based structure applies only to direct customers on the usage-based model. Anyone who joined before February 2024 sits on a separate, older one.
Workato describes the usage component of its pricing as flexible, predictable and scalable. Consumption is weighted, estimated and tracked by capability such as Workflows, APIM and Event Streams. The company says its team helps customers forecast and optimize usage to prevent overconsumption. That is Workato's own characterization of the model, not a measured outcome.
What a buyer can verify directly is the Usage Dashboard. There, customers track task consumption, on-premise agents and concurrency usage, see plan details and configure notifications. For a finance lead, that dashboard is the one place where contracted volume meets real volume.
What the four editions include
Standard is the introductory edition with core integration capabilities. It highlights application integration and process automation, recipe lifecycle management, and dev, test and production environments. The rest of the list: 1 Standard Workbot, SAML SSO and MFA with custom OAuth profiles and role-based access control, encryption at rest and in transit, 30-day log retention and 1 concurrency. If the requirement is a contained set of integrations under basic governance, that is the shape of it.
Business MCP is aimed at broader integration and orchestration needs. It adds pre-built MCP servers, an AI/MCP Gateway, the Enterprise Skill builder, and data integration covering ETL, ELT and reverse ETL. It also brings the API platform, log service, dependency graph and platform APIs, along with 10 Enterprise Workbots, 3 on-premise agents and 1 concurrency.
Enterprise MCP targets large organizations with complex requirements. It adds Data Pipelines, Event Streams, Intelligent Document Processing (IDP), AI by Workato and Insights, plus unlimited Enterprise Workbots. Security deepens here as well: data masking, activity audit, hourly key rotation and runtime user connections. Log retention becomes a flexible 90 days, with 10 on-premise agents and 15 concurrency.
Workato ONE sits at the top of the range. It is the edition that adds agentic capabilities: Agent Studio, Workato Knowledge Base and Workato GO, which covers Employee Assistant, Enterprise Search and Action Board. It carries 30 on-premise agents and 30 concurrency.
Across editions, Workato's own figures, as of September 2026, list 1,200 or more pre-built connectors for every user plus unlimited workspaces and users. Treat the operational details as plan-dependent. Job history, job reports and the ability to re-run jobs all depend on the selected plan.
How usage is counted: business actions and other metrics
Workato bills every capability through one common billing unit, and the counting rules stay consistent across them. Only successful operations count. Steps skipped by conditional logic are excluded. Rerun-eligible operations count. A failed or cancelled job is charged only for the operations completed before the failure.
"Workflow (recipe) usage is calculated in Business actions, which are measured based on the triggers and action steps within a recipe. Not all steps within a recipe automatically count as Business actions."
Source: docs.workato.com
For recipes, the metric is the business action. One is counted for the trigger and for each action-in-an-app step in a recipe. Conditional statements, monitoring blocks and other recipe steps do not count. The usage of a job is the sum of all eligible steps.
Eligible apps include all third-party and community connectors, AI by Workato, HTTP, JDBC, SFTP and code snippets in JavaScript, Python and Ruby. The CSV, JSON and XML tools are on that list as well. Utilities sit outside the meter: Logger, Lookup Tables, Mapper, Message Template, Scheduler, Variables and Data Tables are not counted. Recipe functions and callable recipes inherit the calling recipe's metric.
Other capabilities carry their own meters. The API platform counts API calls, and it counts successful API requests rather than the steps inside API recipes. Managed APIs and the API Gateway convert to credit at different ratios, which the documentation describes without publishing figures. IDP counts Pages processed. Event Streams counts Events processed, where published messages count and consumed messages are free. MCP counts MCP calls. Agent Studio counts Genie actions, where one input prompt equals one Genie action and skill execution and reasoning steps add no usage. Workato GO counts GO seats.
Some of this is additive: IDP pages, Event Streams events and Workflow apps are counted on top of the recipe's primary metric. Usage already counted in one category is never counted again in another.
Add-ons and on-premise connectivity
An edition is not the whole order form. Workato sells optional add-ons on top of it. The list runs: Pre-Packaged Genies, an Advanced Security and Governance add-on, OpenLegacy for on-premise connectivity, RobotiQ RPA, Private Link, on-premises agents, Workflow Apps, Workato Data Hub powered by Reltio, extra concurrencies, Workato EDI powered by Orderful and Virtual Private Workato. None of them carry a public price. The figure comes only in a quote.
On-premise connectivity is worth planning early. Workato supports on-premises applications, databases and custom REST or SOAP applications. Connecting to on-premises systems requires an on-premises gateway. That is a deployment task and a licensing line at the same time, since agents are both an edition allowance and a purchasable add-on.
Two operational settings also live in the contract. Customers can define and configure a data retention policy. Job history, job reports and the ability to re-run jobs depend on the selected plan. Onboarding and training packages are sold separately on request.
What 'Workato cost per task' really means
Workato does not sell tasks. For recipes, the metered unit is the business action. The documentation publishes neither a price per business action nor the credit ratios that different capabilities convert at. Task consumption is the label the Usage Dashboard uses, showing what has been used against the volume in the contract. That is where the search term comes from.
The counting rules are enough to model volume even without a rate. Take a recipe with a trigger and three action-in-an-app steps. That is four business actions per job. Add a Logger step and a conditional branch and it is still four, because neither is eligible. Multiply by real monthly job volume and you have the number that belongs in a quote request.
The rate itself arrives only inside that quote. Any per-task figure on a third-party site is not confirmed by Workato. And since plan structures may vary based on contract terms and join date, someone else's number would not predict yours anyway.
How Workato compares with Zapier, Make, Tray.ai, Boomi and Latenode
| Platform | Billing unit | Published price |
|---|---|---|
| Workato | Business actions | Quote only |
| Zapier | Tasks | From $19.99 a month |
| Make | Credits | Core $9 a month, yearly |
| Tray.ai | Tasks | Quote only |
| Boomi | Subscription or usage | $99 a month plus usage |
| Power Automate | Users and bots | $15 per user, yearly |
| Latenode | CPU seconds | $0.00012 per CPU second |
Zapier publishes a task-based list: Professional from $19.99 a month, Team from $69 with 25 users and SAML SSO, Enterprise by quote. A task is used when a Zap successfully moves data or completes an action. Polling for new data never uses one. With pay-per-task billing on, Zaps keep running past the plan limit at a higher per-task rate. With it off, they pause.
Make publishes credit prices, starting with a free plan of up to 1,000 credits a month. With yearly billing, that is Core at $9 a month, Pro at $16 and Teams at $29, each for 10,000 credits a month. Each module action in a scenario consumes a credit. Make Enterprise adds SSO via OAuth2 or SAML2, role-based access controls, a guaranteed 99.5% uptime SLA and 24/7 priority support with 2-hour response SLAs.
Tray.ai quotes Pro, Team and Enterprise, and measures usage in Tasks that cover integration, automation, MCP and agents. Trials are arranged through its sales team. Boomi sells subscription tiers through sales, plus Pay-As-You-Go at $99 per month plus usage. A 30-day free trial covers the full platform.
MuleSoft lists contact for pricing. Celigo does not publish list prices on its pricing page, checked September 2026. SnapLogic customizes its packages through sales, and Jitterbit requires an annual contract with custom pricing. Power Automate is the enterprise-adjacent exception that does publish. It lists $15.00 per user per month for Premium and $150.00 per bot per month for Process, both paid yearly.
Latenode bills workflow runtime in CPU seconds rather than per operation or per step. The first 10,000 CPU seconds are free every month, then it is $0.00012 per CPU second from 10,001 to 100,000 on Pay as you go. Because the unit is duration, that bill is not directly comparable to tasks, credits or business actions.
So Workato, Tray.ai and MuleSoft price through contracts, while Zapier, Make and Latenode publish self-serve rates. The only fair comparison is total annual cost. Convert one real month of your volume into each unit, then set the quoted platforms against the published ones.
Who Workato's pricing suits and who should look elsewhere
This model fits organizations that already run procurement and staff an integration team. Say you want editions with defined governance, on-premise agents, negotiated annual terms and a vendor that will help forecast usage. In that case, a quoted contract reads as a feature rather than an obstacle.
The watch points are concrete. There is no self-serve checkout, and the pricing page offers a demo and a quote rather than a trial. The total depends on three moving parts: the edition fee, usage and add-ons. Concurrency and on-premise agents are capped per edition, so a workload that needs more parallelism pushes you up a tier or into a paid add-on. And pre-February 2024 customers sit on a different model. That makes peer benchmarks unreliable even when a peer shares real numbers.
Teams that want a price at checkout, a card and a workflow live this afternoon are shopping in a different category. The self-serve platforms publish their rates, so compare those directly. Treat Workato as the option you evaluate when governance and scale justify the sales cycle.
