Software · Verify

Assert — AI-first test automation

Describe what to test in plain language, and your own AI assistant generates a working, version-controlled test in minutes — often with zero hand-written code. Assert is a modern, platform-independent alternative to NI TestStand and LabVIEW, on Linux, macOS and Windows.

Self-hosted — you own your data · tests in open YAML in Git · Linux, macOS & Windows

Why Assert

Test tools built for another era

NI TestStand and LabVIEW are capable tools — but they were built before AI wrote code, before teams lived in Git and before infrastructure ran anywhere. It shows on the bench every day: test logic in binary sequence files an AI can't read and a diff can't show; results as log files you interpret by hand instead of data you can query; and a tie to Windows and proprietary hardware, whether or not it suits your line.

"I've built test systems for 25 years, and the same problems keep coming back: tests locked in files no one can read, results buried in logs, and the knowledge of why walking out the door. Assert is the platform I wish I'd had."
— Daniel Rhodin, founder, E-Sharp AB
From intent to a test — in minutes

Three steps

01

Describe the intent

In plain language: tell your AI assistant what to measure and the limits — e.g. "HDMI output between 4.9 V and 5.1 V". No sequence editor, no boilerplate.

02

AI drafts the test

Your AI code assistant — say Copilot in VS Code — writes it straight as open, readable YAML. Often with zero hand-written code.

03

Review & run

Review the diff, approve and run on any station from the browser. Every measurement is stored as queryable data, tied to the exact test version.

Three parts, one platform

From intent to result

AI-first authoring

Describe what to test in plain language and your own AI assistant drafts the test as readable YAML in Git. Review the diff, approve and run.

Operator UI in the browser

Operators start and monitor tests on any station from a browser — no Windows client to install, no per-seat lock-in. Live status and pass/fail per unit.

Server — data, not log files

Every measurement is written as a structured row in PostgreSQL, not buried in a per-unit log file — so results, Cpk, yield and a live fleet monitor are queryable in SQL.

Data, not log files

Every measurement is a structured, traceable row

Where TestStand writes each run to a report or log file you interpret afterwards, Assert writes every measurement as a structured row in a PostgreSQL database the moment it's taken — with its limits, its verdict and the exact test version attached. Process capability, yield and 3-sigma analysis live in SQL, not in a folder of logs, so outliers and marginal values show before they cost you.

Every result carries its serial number, exact test version, the limits applied and the verdict — so you can trace any measurement back to precisely how and when it was taken. That same structure is what makes the data AI-ready: an assistant can query it, spot drift, correlate failures and explain a result, because every row has meaning.

See Assert on your test bench

Running production or verification tests and want results as queryable data instead of log files? Book a demo and we'll show you what it looks like on your line.

Book a demo of Probe

Show us your board — we review it and go through the findings together.

Prefer to email? info@esharp.se