| Field Name | Description | Required | Notes |
|---|---|---|---|
Base path |
Specifies the URL prefix for all instance paths, relative to the host root. It must start with a leading slash |
Yes |
|
Client provider |
Specifies a client provider for the instance. |
No |
Anypoint Platform acts as the client provider by default. To configure an external client provider, see Client Providers. |
Instance label |
Specifies a label for the instance. |
No |
If you have multiple managed instances of the same asset, add a label to differentiate each instance from the others. |
Advanced Options |
|||
Consumer endpoint |
Specifies a proxy application’s address for consumers to use for sending requests. |
No |
Adding an MCP Bridge Instance
Flex Gateway MCP Bridge enables you to create MCP servers from your existing API instances. With the MCP Bridge, you can expose your current API instances as MCP tools so agents can discover and call them.
Define your MCP bridge by selecting API instances and mapping them to MCP tools. When an agent requests the list of available tools (tools/list), the gateway responds by using an MCP descriptor that defines each tool’s name, description, and input schema. When an agent calls a tool (tools/call), the MCP Bridge maps the tool name and arguments to an HTTP request based on your configuration.
This diagram shows the relationship of the upstream and downstream configurations to your upstream service and agent client:

| MCP Bridge doesn’t support API instances with multiple upstream services. |
Add a New Agent or Tool Instance
To add and a new MCP Bridge for Flex Gateway:
-
Ensure that all API instances you want to add to your MCP Server are running on the same Flex Gateway you want to add your MCP Server to.
-
To begin creating your MCP Server instance, you can start from these locations:
-
From API Instances:
-
Click an API instance you want to add to an MCP Server.
-
Click Actions > Add to new MCP Server instance.
-
-
From Agent and Tool Instances:
-
Click Add > MCP Bridge.
-
-
-
From Select a gateway, choose a Flex Gateway for deploying the MCP server instance.
-
Click Next.
-
Select the API instances you want to add to your MCP Server.
-
Define the MCP server’s asset name.
-
Click Next.
-
Configure the downstream configuration settings:
Managed Flex Gateway
Self-Managed Flex Gateway (Connected Mode)
Field Name Description Required Notes Protocol
Specifies whether to use HTTP or HTTPS.
Yes
If you select HTTPS, specify a TLS context for inbound traffic.
Inbound TLS
TLS Context used for the inbound HTTPS traffic.
No
Configure a TLS Context for Flex Gateway before adding a TLS Context to your server.
Port
Specifies the number to use if the displayed port is incorrect.
No
Instances can share the same port across the same target if the Base path is different for both instances.
Base path
Specifies the URL prefix for all server paths, relative to the host root. It must start with a leading slash
/.Yes
Instance label
Specifies a label for the instance.
No
If you have multiple managed instances of the same asset, add a label to differentiate each instance from the others.
Advanced Options
Consumer endpoint
Specifies a proxy application’s address for consumers to use for sending requests.
No
-
Click Next.
-
Configure the upstream services for your API instances:
Field Name
Description
Required
Notes
Route Label
Label to better identify multiple API instances.
Yes
By default, the label is the API name.
Upstream URL
URL to access for the upstream service. This must end with a
/.Yes
For example, use the URL of your asset in Exchange.
Outbound TLS
TLS Context used for the outbound traffic to the upstream service
No
Configure a TLS Context for Flex Gateway before adding a TLS Context to your server.
To switch among API instances, select the name of the API.
-
Click Next.
-
Define your API Tools:
-
Review your selections and edit them if necessary.
-
If you are ready to deploy, click Save & Deploy. Otherwise, you can select Save, to save the instance and deploy it at a later time.
== Edit Agent Network Project Instances
Some instances in Agent and Tool Instances can be created by supplying a YAML file in an agent network project. For instances created with the YAML file, the YAML file is treated as the source of truth. If you edit the instance in API Manager (for example, to add policies), the changes are reverted to the YAML file when the instance is redeployed from the agent network project.
To ensure changes to instances created from a YAML source are present after redeployment, edit the YAML source.
For more information, see Define Your Agent Network Specification.



