Product Launches
AiEnsured Releases agentest, an Offline Python Tool Running 165 Tests on Agent Traces
By News Desk ·
Key takeaways
- AiEnsured released agentest, an Apache-2.0 Python command-line tool that runs 165 tests on recorded OpenTelemetry GenAI agent traces.
- agentest is deterministic and offline, needs no account, API key or network, and can act as a CI gate through its --strict flag.
- The runner installs from PyPI as agentest-ai, requires Python 3.11 or newer, and ships with a Hugging Face Space and Colab notebook for browser trials.
AiEnsured released agentest, an Apache-2.0 Python command-line tool that runs 165 tests on recorded OpenTelemetry GenAI agent traces, the company said in a PRLog announcement published November 19, 2026. The tool installs from PyPI as agentest-ai and requires Python 3.11 or newer.
What agentest Checks On Recorded Traces
agentest reads traces an agent has already produced, so evaluation needs no live model call. The tool can also read an optional A2A AgentCard and an MCP tools list. Each check returns pass, fail or n/a with a line of evidence.
AiEnsured positions the release for developers and QA engineers who need a repeatable, evidence-backed check on a recorded agent session. Teams can run the 165 tests on a local machine or in a continuous-integration job, and the --strict flag lets a failing check block a build. The tool does not call an LLM judge and needs no account, API key or network.
The open test checklist is published on Hugging Face under CC BY 4.0 and can be used by hand, without the runner. A Hugging Face Space and a Colab notebook are available for a trial in the browser. Dr Srinivas Padmanabhuni is CTO and co-founder of AiEnsured, which operates as testAIng.com.
AiEnsured's Testing Business Behind the Release
AiEnsured launched globally in June 2019 as an AI-focused testing services company, according to akta.pro company records, operating from NOIDA, Bangalore and Jodhpur in India, plus Germany and Canada. The same records list partnerships with Hexaware Technologies, MOH10 Technologies and Infosys for AI testing services, and name MoMAGIC, MyGate, KritiKal, Mobilous, ISRO, BCIC and CII among featured clients. AiEnsured is headquartered in Bengaluru and was founded in 2019.
How Other Vendors Approach Agent Testing
Warp's Ben Holmes demoed LLM judges scoring Claude Code and Codex traces, with his team flagging 100% of redundant tests versus 40% of attendees, StartupHub.ai reported on November 7, 2026. Emtech announced QMT 2.3 on November 18, 2026, extending model-driven testing to legacy and mainframe environments for insurance carriers and Insurtechs, PR.com reported. Meta deployed JiTTests, an LLM-based system that generates tests to fail on risky code changes, and it caught 74 privacy SEVs at the diff stage in H1 2026, MarkTechPost reported on November 16, 2026.