Creating, deploying, and publishing APIs using API Agent
API Agent is designed to accelerate the API journey with greater efficiency, speed, and confidence. This feature is built to work with the rapid evolution of large language models (LLMs) and AI agents, which are driving changes in the world of API consumption and creation.
This new feature enables developers and technical users to create APIs through a chat-based interface, allowing them to define, design, and deploy APIs using natural, conversational interactions.
Overview
API Agent is built on agentic framework so it can plan, understand, and execute user queries. Available as a Visual Studio Code plug-in, embedded in API Manager, and IBM API Studio desktop version. The API Agent simplifies and automates complex API management workflows. By intelligently translating user queries into actions, API Agent streamlines API discovery, creation, governance, and validating.
The API Agent uses an agentic framework that supports a variety of tools to automate tasks such as searching for APIs and tables, and generating code to interact with databases. For a comprehensive list of tools, see Tools compatibility.
| Tools | Description |
|---|---|
| Management MCP | Handles tasks related to IBM® API Connect, such as publishing or securing APIs. |
| Artifacts Searcher | Locates existing resources like APIs, database tables, or application source objects. |
| OAS Creator | Automates the creation of APIs from database tables and IBM connector pack based applications |
| Governance | Provides access to several key API governance capabilities of API Connect through the API Agent. With this tool, you can list rulesets, list rules in rulesets, validate an OpenAPI document using rulesets, and fix an OpenAPI document with supported rulesets. |
| OpenAPI Python Generator | Generates boilerplate or custom code from OpenAPI specifications using the FastAPI framework. |
| OpenAPI Enhancer | Completes upgrading or modifying an input OpenAPI file or API with enhanced descriptions and examples, and automatically rectifies any OpenAPI file with validation problems or errors. |
| OpenAPI Generator | Generates OpenAPI documents from a natural language description (NLP). The user
provides a prompt to the API Agent, describing the
intended functionality of the application. Based on this prompt, the API Agent generates an
OpenAPI specification that represents the described API. The framework ensures that tasks are routed to the appropriate tool based on the user's query. In the initial version, API Agent confirms with the user whether the proposed plan of action is accurate before executing it, ensuring precision and user control. |
| Code deployment | Streamlines the deployment of application source code to supported services, providing a flexible, and controlled way to manage deployments. |
| Documentation searcher | Searches knowledge center or user-uploaded documents. |
| Analytics | Query API Connect analytics using natural language and visualize insights with AI-generated charts. |
| API discovery and rationalization | Identifies duplicate or similar APIs, helping you organize catalogs more effectively. |
| MCP server and client | Creates servers and registers or uses MCP servers as tools. |
| Governance | Validate the OpenAPI documents against spectral rulesets. |
| Governance MCP | Validate spectral rules against OpenAPI definitions, create rulesets, and update existing rulesets |
| Management AI MCP | Generate and modify gateway policies and create spectral rules from natural language descriptions. |
| IBM API Studio | Publishes projects, creates test cases, runs test cases, and creates NLP-based gateway policies. |
| Help | Get interactive, task-specific guidance and contextual help for using the API Agent and its tools. |
API Agent components
API Agent has the following components:
-
- API Agent sources
- The intelligence behind API Agent is powered by
API Agent sources configured by
administrators to integrate various sources that fuel the system's capabilities. Leveraging a robust
Retrieval-Augmented Generation (RAG) pattern, this catalog crawls and indexes supported sources,
including APIs and database tables, storing them as vector embeddings. Key capabilities:
- Acting as a centralized hub for managing and configuring the resources required for API development.
- Enabling seamless connection and configuration of data sources like databases, and APIs.
- Providing immediate access to all configured resources for creating and managing APIs.
-
- Chat interface
- Using the Visual Studio Code, API Manager, and IBM API Studio desktop version you can
use API Agent that
provides a chat interface through which you can enter different prompts. For example,
Search for available APIsorCreate an openapi with the table purchaseorder.The API Agent connects to API Agent sources, retrieves the necessary information, and provides actionable results.
The following topics provide full details for configuring and using API Agent.