Visit https://docs.parasoft.com/display/SVC20262/Release+Notes for the latest release information.

This release includes the following enhancements.

Release date: October 28, 2026

Agentic API Testing 

SOAtest 2026.2 brings API testing into your agentic workflows with new MCP tools and AI agent skills that make SOAtest testing capabilities available to your AI agents. You can create, run, and manage API tests through an AI agent or LLM client such as GitHub Copilot, then save the resulting scenarios as persistent, deterministic SOAtest assets. 

You can give your AI agent requirements, acceptance criteria, and API definitions as context for generating API test scenarios. Your agent can also run tests, review the results, and update the scenarios based on what it finds. Included AI agent skills guide you through API test creation and provide scenario execution capabilities. 

With the agentic testing workflow, you can: 

• Create, modify, describe, and delete test scenarios and individual tests 
• Build tests for REST APIs, SOAP services, and other HTTP endpoints 
• Work with JSON, XML, GraphQL, and other text-based message formats 
• Automatically create data banks that extract response data so you can reuse it in later requests 
• Automatically create JSON and XML assertions to validate responses 
• Apply Smart Test Templates during test creation so you can reuse predefined testing configurations, including common authentication patterns 
• Create, modify, and manage CSV and Excel data sources 
• Make HTTP requests directly through MCP tools as part of your agent-driven workflows 

These capabilities let you automate significant portions of API test design and execution while continuing to use proven SOAtest testing workflows. 

Agentic Stateful Virtual Services Creation 

Virtualize 2026.2 lets you create stateful virtual services through new MCP-enabled workflows and AI skills. You can use an AI agent or LLM client such as GitHub Copilot to generate virtual services that maintain state across interactions. This means earlier requests can influence the data returned by later requests, giving you a more realistic simulation. You can model both happy-path and error-handling scenarios and create stateful assets through agentic workflows without needing to use the Virtualize desktop environment. 

To support stateful behavior, Virtualize provides a new document repository to store service state, data repository data source integration to initialize document repository content, and a Document Repository CRUD tool to configure scripts that read and update stored state. A dedicated AI skill coordinates the end-to-end creation process through MCP, helping you build complex stateful virtual assets more easily. 

Improved Agentic Stateless Virtual Services Creation 

Virtualize 2026.2 gives you a more reliable AI-driven workflow for creating stateless virtual services. A new AI skill coordinates the creation process, so you can reduce manual configuration and generate virtual assets more consistently through agentic workflows. 

The updated workflow automatically configures inline expressions in response payloads to parameterize data from incoming requests. This helps you create flexible, realistic virtual services without extra manual customization. Your LLM client no longer needs to support MCP sampling to populate data, so you can use a wider range of AI agents and LLM clients. 

Together, these enhancements help you generate and manage stateless virtual services more effectively while producing virtual assets that you can maintain and extend through your existing Virtualize workflows. 

Integration Library for Coverage and TIA workflows 

A new Java-based integration library is now available that helps you collect code coverage and run Test Impact Analysis (TIA) when you use third-party test frameworks such as JUnit, TestNG, Playwright, Selenium, and Cucumber. You can add coverage and TIA capabilities to your existing test automation pipelines with less integration work. The library manages communication with coverage agents, propagates the required headers to your application under test, and handles concurrent test execution for you. The library is available here. 

Enhanced Reporting for Projects with Partial Automated Functional Test Runs 

CTP 2026.1 introduced code coverage reporting and test impact analysis for partial automated functional test runs. This release extends those capabilities with enhanced reporting in Parasoft DTP for those workflows.

Enhanced Precision in Test Impact Analysis 

Method-level granularity is now supported in all Test Impact Analysis (TIA) workflows to identify tests affected by code changes more precisely. This includes workflows that use application binaries to identify application source code and workflows that use cumulative builds in DTP to aggregate reporting for projects with partial automated functional test runs.

Breaking Changes

• The Parasoft Data Repository Server available for download and embedded within the desktop installation is no longer supported and has been removed. You must replace it with your own MongoDB installation and migrate data to it to continue running a data repository server. 
  • If you have data in a Parasoft data repository that you want to keep, a migration utility is provided on the Parasoft Marketplace. See Migrating a Data Repository Server to a MongoDB Server for more information about using the utility to migrate your Parasoft Data Repository Server to your own MongoDB Server. You can do the migration before or after upgrading to 2026.2. 
• Support for Tomcat 10.1 has ended. You must upgrade to Tomcat 11 before updating CTP or the SOAtest and Virtualize WAR deployment. See Apache Tomcat Example Deployment for information about installing Tomcat for WAR deployments. After installation, you can copy your config.properties file and your workspace from your old deployment into the new one. See Upgrading for Apache Tomcat Deployments for information about copying the file from one installation to another. 
• Event monitoring and hit statistics workflows from CTP, Virtualize desktop, and the SOAtest Event Monitor tool now require Java 21. You must upgrade CTP or the SOAtest and Virtualize WAR deployment to use Java 21 if you use event monitoring or hit statistics workflows. SOAtest and Virtualize desktop already ships with Java 21. 
• CTP users who use DTP cumulative builds for aggregating data from multiple partial automated functional test runs must upgrade both CTP and DTP to version 2026.2 at the same time to avoid compatibility issues. 
• If your team uses coverage workflows, you must also upgrade your Java coverage agents to version 2026.2. The data format for the Java coverage agents has changed and data collected by existing coverage agents is incompatible with CTP and SOAtest 2026.2. 

Additional Updates

• You can now configure the performance settings for a message responder using the REST API. 
• The Diff tool can now resolve ${} variables in external control files, so now you can diff against expected dynamic data without having to save the control source in the internal editor.  See Diff Tool. 
• You can now configure a Smart Test Template to add a client keystore for 2-way SSL. See Creating Smart Test Templates. 
• GitHub Copilot now comes installed in SOAtest desktop. 
• SOAtest desktop now ships with Tomcat 11. 
• The gRPC transport and listener now support client certificates for 2-way SSL. 
• Apache Avro Client can now access a Schema Registry that requires authentication. 
• Support for Java 17 is deprecated and will be removed in a future release. 
• The AI Assistant is deprecated and will be removed in a future release. 
• The Parabank demo project bundled with SOAtest desktop is deprecated and will be removed in a future release. Parabank should be deployed from Docker Hub or GitHub instead. 

Resolved PRs and FRs

SOAtest and Virtualize

CTP