Describe a user flow.
BuniOD tests it in a real browser.

BuniOD is an AI test automation platform for web applications, made by the software company NewProWeb. You write a user flow as plain-language test steps with the result you expect, and BuniOD’s AI QA agent turns them into a browser scenario with test data and checks. A person on your team confirms the scenario, then it runs in a real browser on demand or once a day, and every run keeps screenshots, logs and an explanation of any failure.

Playwright is a code framework: engineers write and maintain the test scripts. With BuniOD you write the steps in plain language, review the scenario it builds and run it in a real browser — there is no test code to maintain. mabl and testRigor are AI-assisted tools too, and today they cover more browsers, mobile apps and CI integrations. BuniOD is simpler to start with: a Free plan, paid plans from $39/mo and unlimited projects and scenarios on every plan. See the sourced side-by-side comparison.

No. You write the steps and the expected result in plain language, and BuniOD matches each step to your pages and adds the checks. You can adjust input values or regenerate the scenario from an updated prompt.

No. BuniOD automates the repetitive part — re-clicking the same test cases every release. Deciding what to test, exploratory testing and judging results stay with people, and every generated scenario is reviewed by a person before it runs.

Pricing is based on AI tokens — the units of AI work used to generate and run your tests. Free is $0 with 500K tokens a month, Pro is $39/mo with 5M and Scale is $59/mo with 20M. Every plan includes unlimited projects and scenarios. On Free, runs pause when the tokens run out; on Pro, work continues and extra tokens cost $3.99 per 1M. Full pricing.

On day one: connect your URL, write the steps of one flow, review the scenario and run it.

Yes. Public pages need only the URL. For signed-in areas, add HTTP basic auth (stored encrypted) or a test identity with its own inbox, which also lets scenarios check emails your app sends. Use a dedicated test account and a non-production environment where possible, because scenarios can create or change data.

The run stops at the first failed step or check and records which one failed, an explanation, a screenshot and console and network logs. Scenarios that depend on it in a business process are skipped. Failed runs are not retried automatically.

Today BuniOD sends in-app and email notifications when runs finish or fail and when the agent has a question. Each person chooses which emails they receive. Jira and GitHub Actions integrations are coming soon.

CAPTCHA. Scenarios act like a real user, so a CAPTCHA can stop a run — point BuniOD at a test environment with CAPTCHA disabled or a test key. Runs also use a desktop Chromium browser, so native mobile apps are not covered today.

Google Gemini models generate scenarios and OpenAI GPT-5 drives runs, with automatic fallback to another model if one fails. Your prompts and the structure of tested pages are processed by these providers.

Neither. You don’t write code: you write the steps in plain language, BuniOD builds a browser scenario from them, you review and confirm it, and it runs in a real browser. A browser automation engine runs underneath, but you work with scenarios, not scripts.

Both, for web applications. Each scenario drives the real interface of your web app in a browser, so it works as automated UI testing, and it follows a whole user flow from start to finish, so it is also end-to-end testing. It is not a unit, API or load testing tool.