Frequently Asked Questions

Getting started

Install API Agent

Adding the Visual Studio Code plug-in, setting up standalone IBM API Studio, or use IBM API Studio in API Manager. You can choose any of the options based on your preference.

Do I need to configure anything before using API Agent in IBM API Studio?

You must configure API Manager before using API Agent.

What can I do after installing

You can do the following actions:

• Generate OpenAPI specifications
• Run tests
• Create code
• Interact with API Connect using natural language prompts
• Automate tasks such as validating APIs and managing artifacts across supported environments

How does this help me?

It simplifies API development and governance by reducing manual effort through prompt-based automation.

How do I know which tools support which prompt actions?

A compatibility table shows which tools (Visual Studio Code, API Manager, IBM API Studio) support specific prompt-based actions. See Tools compatibility.

API Agent roles and permissions

User roles define what actions a user can perform in the API Agent. Roles include combinations of permissions that control access to features and configuration options.

What is the Api-Agent permission used for?

The Api-Agent permission enables users to use the Visual Studio Code plug-in ,IBM API Studio, and API Manager.

Which roles include the Api-Agent permission?

API Agent User: Includes Api-Agent and all Viewer permissions. Owner and Administrator: Include all permissions, including Api-Agent.

What can an Owner or Administrator do?

Users with these roles can configure the system, use the Visual Studio Code plug-in, IBM API Studio, and API Manager to access all available permissions.

What does the API Agent User role allow?

This role allows users to use the Visual Studio Code plug-in, IBM API Studio, and API Manager to complete all actions permitted to a Viewer.

What permissions are available in the UI?

Settings: View and manage configuration settings. API Agent: Use the Visual Studio Code plug-in, IBM API Studio, and API Manager.

What is the Viewer role?

The Viewer role typically includes read-only access to view configurations and data without making changes.

Settings

Watsonx.ai Settings

You need a valid IBM Cloud API key or bearer token with permissions to access foundation models and a project ID or deployment space ID linked to your Watsonx.ai service.

What happens if I don’t provide credentials?

For SaaS environments, IBM-managed credentials are used by default if none are provided.

What should I do before configuring Watsonx.ai in an on-premises setup?

You must first enable the API Agent on the management CR.

Catalogs and Drafts

The legacy Drafts and Sandbox catalogs are automatically added once the API Agent is enabled.

Can I add more catalogs to the API Agent?

Yes. You can add additional catalogs from your provider organization using the Settings > API Agent > Sources menu.

How do I view the updated catalog list?

After saving changes, expand the provider organization list to view the updated catalog list.

What permissions do I need to manage application sources?

You must have administrator and API Agent user permissions to view the API Agent tab and access the API Agent sources.

What actions can I perform on application sources?

You can do the following actions:

1. Add a new application source
2. Edit the object list
3. Refresh application sources manually
4. Configure auto refresh
5. Edit connection details
6. Remove application sources

What happens when I remove an application source?

Removing an application source deletes it from the UI and backend, making its data inaccessible.

What authentication method is supported for MCP servers?

Currently, only token-based authentication is supported. Enter the token details in the Authorization dialog box and click Test connection.

How do I change the list of tool operations?

Administrators can update the list of tool operations only from the Tools page in API Manager. To know more details about how to update list of tools operation, see Configuring MCP tools.

Best practices

Why should the extension window occupy full height and at least one-third of the Visual Studio Code window?

This layout ensures optimal performance and proper rendering of the API Agent interface. A smaller window can cause incomplete display of prompts and results.

What should I do if the system does not understand my request?

The system provides alternative suggestions when it cannot interpret your prompt. Review these suggestions and rephrase your request with clearer context and keywords.

What should I do if answers differ from official documentation?

Always cross-check critical details with official IBM documentation before implementing changes. This prevents errors caused by outdated or incomplete responses.

Using API Agent

What is the primary purpose of the API Agent plug-in?

The API Agent helps manage the API lifecycle by enabling tasks such as searching metadata, generating OpenAPI documents, creating and deploying Python code, and validating APIs before publishing.

What filtering options are available during API search?

You can filter results by OpenAPI versions and API types to narrow down the search.

How do I generate an OpenAPI document from an existing source?

Enter a prompt like:

Create an openapi with the table orders

Create an openapi with the table orders and schema purchase_order

Note: The schema name is optional unless the table name is not unique.

What file format is required for deploying an application to IBM Code Engine?

The application code must be packaged as a compressed file generated by the OpenAPI to Python tool. Other formats are not supported. For more information, see Code deployment.

Tools

What is the purpose of different tools?

The purpose of API Agent tools is to provide a programmatic interface for interacting with the API Agent, enabling automation and integration with other tools and systems.

Where can I get list of all tools?

See Tools compatibility to know more details about each tools.