Wendell Runner is an open-source command-line tool designed for regression testing of agent workflows, focusing on catching policy and business logic regressions that traditional unit or API tests may overlook. It addresses scenarios where changes to agent prompts or workflows can lead to unintended behaviors, such as unauthorized refunds or missed escalations, by allowing developers to encode and test these critical paths as committed regression tests.
The tool operates locally or within continuous integration environments, requiring no hosted dashboard or account. Developers define readable JSON scenarios alongside their agent code, prompts, policies, and fixtures. Wendell Runner executes these scenarios using an agent command that can invoke any process capable of reading JSON from standard input and producing results to standard output, supporting runtimes like Python, TypeScript, and Node. Tests are configured via a simple configuration file, and the tool can assert the presence or order of tool calls, message contents, and other workflow-specific behaviors.
Wendell Runner integrates with CI pipelines by providing real exit codes and generating reports in multiple formats, including JSON, JUnit, and GitHub summaries. This enables both human-readable terminal output and structured results for automation or pull request annotations. The tool is intended for developers who want to ensure that agent-driven workflows comply with business rules and policies, especially in contexts where prompt or code changes could introduce subtle regressions.
The platform is distributed as a PyPI package and can be installed with tools like uv or pipx, without requiring a git checkout. Wendell Runner is released under the MIT license, with its public source code and releases available on GitHub. Its local-first, open-source approach makes it suitable for teams seeking CI-ready agent regression testing without external dependencies.
In the LLM eval & observability space, Wendell Ai takes a focused approach. It focuses on ensuring AI agent workflows behave as intended before deploying to production environments. Wendell Ai is a B2B product aimed at AI developers and automation engineers. The product ships for the command line.
Wendell Ai first shipped in 2026. Across PulseGate's embedding index, Wendell Ai has few near neighbours, marking it as relatively distinct. Among its 5 catalogued features are playbook execution, workflow validation, and scenario reporting.
Latest indexed changes and source events
wendellai.com discovered by the PulseGate indexer
Other apps tracked under the same category.
