IronBee Express
Goal-driven browser agent where each step is one Jev choice over the page's controls, then reviews the run for root causes.
Overview
IronBee Express is a browser agent from IronBee for checking that a web app really did what the page says. You describe a task in one sentence, such as log in, add an item to the cart and open the cart, and it carries it out in a real browser. Each step is one TypeSafe Jev decision over the page's offered controls plus one call to IronBee DevTools, which performs the action and returns the next snapshot. Model output never becomes a selector, coordinate or script.
After the run it reviews the final page, API responses and, with the IronBee platform, backend logs and traces, so a success message alongside a PAYMENT_FAILED response fails the run and points to the cause. There is no LLM in the loop by default; a text model is optional for deeper reasoning. The README reports about 300 ms per step, a 9-action run in 6.7 s costing $0.00054 including review, and a 20-action demo in about 12 s, while recorded scenarios replay in about 4 s with no step decisions. It is Elastic-2.0 licensed with a waitlist for the hosted product.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Which of the page's controls, options or values should be used next to pursue the goal? | — |
| Yes / no | Did the app really do what the page claims?(inferred) | yes · no |
How it works
- User states a one-sentence goal
- Snapshot the page's controls via IronBee DevTools
- Jev chooses one offered control per step
- Execute and repeat until done; record the scenario
- Review page, API responses and backend traces for a pass/fail verdict
Strengths
- Actions can only target offered controls, never free-form selectors
- README reports about 300 ms per step and sub-cent run costs
- Recorded runs replay with no decisions
- Root-cause review across frontend and backend
Caveats
- Elastic-2.0 licence restricts some commercial uses
- Full backend review needs the IronBee platform
- Hosted product is waitlisted
At a glance
- Type
- App
- Status
- beta
- Language
- TypeScript
- Runtime
- Node CLI with a real browser
- Has code
- yes
- License
- Elastic-2.0
Use cases and integrations
How curated lists describe it
Browser agent for checking that a web app really did what the page says: each step is one Jev choice over the page's controls, a text model is optional, and a recorded run replays with no step decisions. Elastic-2.0.
MrJev/awesome-jev
Frequently asked questions
What does Jev decide in IronBee Express?
Which offered page control to act on at each step toward the goal.
Is an LLM required?
No. Jev decides every step by default; a text model is optional.
Do I write assertions?
No. The run is reviewed automatically against page state, API responses and, with IronBee, backend logs.
Related projects
Browser agent with no LLM: code enumerates options from the page and goal, and Jev only picks actions, values and when to stop.
ProjectprototypePythonTypeScript library and CLI for exploratory browser testing where Jev chooses actions and judges results and Playwright executes them.
ProjectprototypeTypeScriptCodexQA browser skill that indexes in-page controls and has Jev pick the operation and target each step, with no screenshots needed.
Agent skillbetaChinese UI test framework where Jev scores candidate page elements for each step of a natural-language test case.
ProjectbetaPythonCucumber runner where Jev resolves each Gherkin step against the live page and Playwright executes it, so tests need only the .feature file.
ProjectbetaTypeScriptFlutter integration-test package where Jev picks the next registered UI action and judges whether the test goal is reached.
LibrarybetaDart