Skip to content

API and MCP enablement

InfoConnect Hub surfaces IBM i business logic and data as REST APIs without changes to your IBM i applications. The Hub also provides a ready-to-use, secure MCP server that surfaces the same APIs as tools for AI agents.

API types

API type What it surfaces Guides
Program Call An IBM i program or service program. Program Call, API parameters and validation
Data API An SQL statement or a stored procedure that runs on IBM i. Data API, Data API configuration
Screen automation (RPA Call) A sequence of green-screen actions. RPA Call, RPA Hub user guide, RPA Hub API reference

Common tasks

MCP server and AI agents

The MCP server generates a tool from each Program Call, Data API and RPA Call stored in the Hub, and you manage these tools at run time. AI agents and chatbots use the tools to discover and run your APIs in AI workflows. See the AI and MCP tutorial, which also describes the MCP Explorer in the web UI.

Admin API

Use the Admin API to set up the Hub manually (with cURL, Postman or another API client) or as part of a CI/CD flow. It provides the following functions:

  • Encryption API: encrypts sensitive data such as passwords and other secrets.
  • Connections: manages IBM i connection definitions. One Hub instance can connect to several back-end systems. In most cases a dedicated one-to-one relationship works best, because it keeps the access rules separate.
  • Program Calls: manages the Program Call definitions that the Hub provides to consumers as REST APIs.
  • Data APIs: manages API definitions based on SQL statements and stored procedures that run on IBM i. See the Data API Configuration Guide.
  • Data Queue Sources (outbound Data Flows): always-on DQ listeners that receive new IBM i data queue messages immediately and send them in real time to these targets:
    • Google Cloud Pub/Sub
    • JMS (ActiveMQ, ArtemisMQ, AmazonMQ)
    • Azure (Service Bus, Event Hubs, Event Grid)
    • AWS (SQS, SNS)
    • Kafka
    • JDBC
    • Files on Amazon S3, Azure Data Lake Storage (ADLS), Azure Blob Storage, Google Cloud Storage (GCS), FTP and SFTP
  • Data Queue Sinks (inbound Data Flows): always-on listeners of messaging and event platforms that send new data to IBM i data queues in real time. The supported platforms are:
    • Google Cloud Pub/Sub
    • JMS (ActiveMQ, ArtemisMQ, AmazonMQ)
    • Azure (Service Bus, Event Hubs)
    • AWS (SQS)
    • Kafka
  • User Groups: manages users, groups and API permissions. See User Groups and Permissions.
  • OpenAPI generation: generates OpenAPI specifications for the defined APIs. See Managing Hub APIs.

Program Calls and DQ pollers support JSONata transformations for input and output payloads. For details, see the API Reference.

In the web interface, Data Queue Sources and Data Queue Sinks are Data Flows. In the Admin API and in JSON they are DQ pollers (/admin/dqpollers): pollerType Source is outbound and Sink is inbound.

Functional API

The Hub provides these Functional APIs:

  • Execute Command Call: needs no configuration.
  • Publish Data To Data Queue and read a Data Queue entry: need no configuration.
  • Execute Program Call: needs a Program Call defined through the Admin API.
  • DQ Listener: listens for new data queue entries (needs a configuration through the Admin API) and sends them to the targets listed under Data Queue Sources.
  • Execute Data API: needs a Data API defined through the Admin API.
  • Execute Screen Automation script.

For details, see the API Reference.