--- title: "MuleSoft Anypoint Integration" description: "MuleSoft Anypoint Platform is Salesforce's integration platform. Workflows search Exchange for assets, list environments, start, stop, and update CloudHub applications, read their logs and deployments, and inspect managed APIs." url: https://flowrunner.ai/integrations/mulesoft date_modified: 2026-09-04T14:59:52-07:00 --- # MuleSoft Anypoint [Developer Tools](https://flowrunner.ai/integrations/category/developer-infrastructure) MuleSoft Anypoint Platform is Salesforce's integration platform. Workflows search Exchange for assets, list environments, start, stop, and update CloudHub applications, read their logs and deployments, and inspect managed APIs. 15 actions · OAuth · available [MuleSoft Anypoint website](https://anypoint.mulesoft.com/) · [Platform Documentation](https://docs.mulesoft.com/) · Capability data verified 2026-08-25 1. A scheduled sweep runs across every Anypoint environment the connection can reach 2. List Applications and Get Application report each CloudHub deployment's state 3. An application reporting STARTED with a failing worker is flagged, because those two states diverge 4. Get Application Deployments resolves the deployment id, and Get Application Logs pulls the stack trace 5. A summary reaches the on-call channel with the failing worker and the log excerpt 6. A ticket opens in the tracker with the deployment history attached 7. The integration owner decides whether to restart the application or leave it running degraded ## What This Integration Enables Anypoint Platform is where a large company's integrations already run. That is a fact worth taking seriously rather than arguing with. The APIs, the CloudHub applications, the policies in API Manager, and the contracts in Exchange represent years of committed work, and none of it is going anywhere because a newer tool arrived. FlowRunner's position here is not replacement. It is that an integration platform is itself a production system, and production systems need operating by more people than the ones who hold logins to them. FlowRunner [connectors](https://flowrunner.ai/concepts/connectors) reach four separate Anypoint APIs that share a hostname and nothing else: Access Management for the organization and environment ids everything depends on, Exchange for the asset catalog, Runtime Manager for CloudHub deployments and their logs, and API Manager for the policy layer in front of them. Agents watch deployments, gather evidence when one fails, check a domain before a deployment claims it, and start or stop applications on a schedule. What they do not do unattended is change a running integration's configuration, because on CloudHub that is not an edit. That is [human-in-the-loop](https://flowrunner.ai/concepts/human-in-the-loop) territory, and the reason is in the next section. ### Without FlowRunner **Platform access as the bottleneck**: Only a handful of people hold Anypoint credentials, so everyone else waits on them for a status answer **Green until it is not**: An application reports STARTED while one of its workers fails, and nothing surfaces the difference **Logs behind three lookups**: The stack trace lives under a deployment id that nobody can find without opening Runtime Manager ### With FlowRunner **Status where the team already is**: Deployment state, worker health, and the relevant log arrive in the channel that handles the incident **The divergence is the alert**: Agents compare the deployment's state against each worker's, and flag the gap rather than the headline **Evidence gathered before anyone asks**: The deployment history and the failing deployment's log are attached to the ticket at creation ## Use Case Scenarios ### The failure explains itself before anyone asks An application stops reporting healthy. The agent calls Get Application and reads not just the deployment status but each entry in the worker statuses, because a deployment can sit at STARTED while one worker fails underneath it. It calls Get Application Deployments to find the id of the deployment that actually failed, then Get Application Logs against that id, which is the only place the stack trace appears. The summary posts to [Slack](https://flowrunner.ai/integrations/slack) or [Microsoft Teams](https://flowrunner.ai/integrations/microsoft-teams), a ticket opens in [ServiceNow](https://flowrunner.ai/integrations/servicenow) with the log attached, and an alert raises in [Datadog](https://flowrunner.ai/integrations/datadog) tagged with the environment. Nobody has opened Runtime Manager yet, and everybody already knows what broke. ### Naming a deployment before it fails at the end A CloudHub domain is unique across all of CloudHub, not just your organization, which means a plausible name for a new integration is very often already taken by a stranger. A deployment that discovers this discovers it at the end. So before any provisioning workflow proceeds, the agent calls Check Domain Available on the candidate name and its fallbacks, and only the free one goes forward. The same workflow calls Search Exchange Assets first to see whether a connector or template for the target system already exists in the catalog, which is a public route and answers before a Connected App is even configured. ### Scheduled quiet hours with the production flag honored A cost workflow stops non-production applications outside working hours with Set Application Status and starts them again in the morning. The interesting part is the guard. The agent calls List Environments and reads the production flag on each one rather than trusting the environment's name, because an environment called Staging can be flagged production and be governed and billed accordingly. Anything carrying that flag is excluded from the schedule and reported instead. The agent also polls Get Application afterwards rather than assuming, since a status change is accepted immediately and takes minutes to take effect. ## Human-in-Loop Highlight Update Application reads like a settings change and behaves like a deployment. CloudHub redeploys the application to apply most of what this operation changes, so adjusting a worker size or a runtime version takes the integration offline while it comes back. Worse, the properties field replaces the entire property map rather than merging into it, which means an agent sending one new property wipes every other property the application was configured with. So a configuration workflow never writes directly. It reads the application first, computes the merged map, and posts the difference: "Applying the new queue endpoint to orders-api in Production. This send includes 14 existing properties and adds 1. Three properties currently set would be removed by this call. The environment is flagged production and CloudHub will redeploy to apply it, so the integration is offline during the restart. Proceed?" The owner reads the removal list and answers. The agent did the merge, the comparison, and the environment check. The person owned the outage. Agent processes routinely Detects exception requiring judgment Clear match Continues automatically Ambiguous Routes to human via preferred channel Human decides Agent resumes with decision ## Agent Capabilities 15 actions ### Access Management 3 - **Get My Profile** Returns the identity the credentials resolve to along with the organization tree. The organization id every other Anypoint API needs comes from here and nowhere else. - **List Organizations** Returns the organizations and business groups the credentials can reach. A nested business group is a separate organization id, and a Connected App scoped only to the root cannot see what is deployed inside one. - **List Environments** Returns an organization's environments with their ids. Every Runtime Manager call needs one, and the production flag on each matters more than its name. ### Exchange 2 - **Search Exchange Assets** Searches the Anypoint Exchange catalog for connectors, APIs, templates, and examples. This route answers without credentials, and supplying them additionally surfaces your organization's private assets. - **Get Exchange Asset** Returns one Exchange asset in full. An asset with many versions returns a very large response, because every version's file list is included. ### Runtime Manager and CloudHub 8 - **List Applications** Returns the applications deployed to CloudHub in one environment. The environment travels in a header, so the same call against a different environment returns a different list. - **Get Application** Returns one application's configuration and state. Read the individual worker statuses as well as the deployment status, because an application can report STARTED with a worker failing. - **Set Application Status** Starts, stops, or restarts a deployed application. Stopping takes the integration offline immediately, and CloudHub still bills its workers while it is stopped. - **Update Application** Changes a deployed application's worker count and size, runtime version, region, and properties. Most changes trigger a redeploy, and the property map is replaced rather than merged. - **Delete Application** Undeploys and deletes an application. Irreversible, and it releases the CloudHub domain name for anyone on the platform to claim. - **Get Application Logs** Returns log records for a specific deployment. The only place the actual stack trace from a failed deployment appears. - **Get Application Deployments** Returns an application's deployment history, and the source of the deployment id that Get Application Logs requires. - **Check Domain Available** Reports whether a CloudHub domain name is free. Worth calling before any provisioning workflow, because the namespace is global. ### API Manager 2 - **List Managed APIs** Returns the APIs registered in API Manager for an environment. A managed API is the policy and contract layer in front of a deployment, and it is a separate record from the application itself. - **Get Managed API** Returns one API Manager instance with its endpoint and policies. The identifier is the API instance id, not the Exchange asset id, since one asset can have several managed instances. ## Frequently Asked Questions ### What can FlowRunner do with MuleSoft Anypoint? FlowRunner agents can run Get My Profile, List Organizations, and List Environments in MuleSoft Anypoint, plus 12 more actions. ### Does connecting MuleSoft Anypoint to FlowRunner require OAuth? Yes. MuleSoft Anypoint connects to FlowRunner with OAuth 2.0, so agents authenticate without handling raw credentials. ### Can MuleSoft Anypoint trigger a FlowRunner workflow automatically? MuleSoft Anypoint doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger. **Work at MuleSoft Anypoint?** This integration exposes MuleSoft Anypoint to AI agents on every FlowRunner plan, including through MCP, at no cost to you. [See what FlowRunner offers integration partners](https://flowrunner.ai/integrations/partners), including how to keep this page current. --- Markdown version of https://flowrunner.ai/integrations/mulesoft. Site index: https://flowrunner.ai/llms.txt