Selenium, Cypress, Playwright and Puppeteer are excellent libraries. They are also projects you have to write, host, schedule and repair. This page is an honest side-by-side, including the places where the framework wins.
It is free, familiar and genuinely good at what it does: open the page, click around, look at what is there. Then the session ends and nothing remains. No saved scenario, no report, no history, no schedule. ReTibe now looks at the page too before it writes a step, so the difference is no longer who can see the page. It is what is left afterwards.
| After the session | Playwright MCP | ReTibe MCP |
|---|---|---|
| What is left | Nothing | A saved scenario, a report, run history |
| Running it again | Explain the whole flow again | One call by id |
| Looks at the page before writing | Yes, that is the point | Yes: a page outline with selectors verified unique, and a site crawl |
| On a schedule | Only while you are sitting there | Cron, daily, weekly; email or a signed webhook on failure |
| A test that passes and checks nothing | Nobody notices | Validator and storyboard flag it before a browser opens, without spending a run |
| Telling a flake from a regression | Nothing to compare against | Per-step outcomes across runs, and a two-run diff |
| When a step fails | Whatever the model happened to see | Its screenshot, console errors and failed requests, kept by default |
They are not competitors, and our own docs say so. Playwright MCP is still the tool for driving a browser by hand, and ReTibe no longer needs it to see the page. Keep it if you like it, and hand the flow to ReTibe so it keeps running after you close the laptop.
Our free plan runs 20 web tests a month, forever, plus 60 page inspections on a counter of their own. The cost of the habit is not money, it is the regression nobody was watching for.
Add our MCP server next to the one you already have and ask the same agent to pin one flow.
See the connectorWhere a framework needs a library, a plugin or a paid cloud add-on, we say so.
| Feature | ReTibe | Selenium | Cypress | Playwright | Puppeteer |
|---|---|---|---|---|---|
| Web interface, no code | JSON + visual | Code | Code | Code | Code |
| Automatic crawling | Built-in | No | No | No | No |
| Visual regression | Baseline / diff | Libraries | Plugins | Built-in basic | Libraries |
| Security scanner | 17 categories | No | No | No | No |
| SEO analysis | Links, sitemap, robots | No | No | No | No |
| Multi-browser | 3 engines | All major | Chrome, Edge | 3 engines | Chromium only |
| Mobile emulation | 16 devices, touch | Partial | Viewport only | Devices | Viewport only |
| Accessibility (a11y) | axe-core + Lighthouse | Libraries | Plugins | Libraries | Libraries |
| Performance analysis | Core Web Vitals | No | Plugins | Traces | Manual |
| Network / CPU throttling | 5 profiles + custom | No | Limited | Yes | Yes |
| Memory profiling | Heap, nodes, listeners | No | No | Manual | Manual |
| Load testing lite | RPS, P95, error rate | No | No | No | No |
| Data-driven testing | CSV, JSON, inline | Code | Fixtures | Code | Code |
| Testing calendar | Built-in | External CI | External CI | External CI | External CI |
| Run history and analytics | Timeline, heatmap, flaky vs regressed | No | Paid cloud | HTML report | No |
| Recording mode | Browser to JSON | IDE recorder | Studio | Codegen | No |
| Figma comparison | Pixel-perfect diff | No | No | No | No |
| Jira issue on failure | Bug with screenshots | No | No | No | No |
| AI assistant | Chat, voice, report analysis | No | No | No | No |
| MCP server for coding agents | 28 tools (26 hosted), 3 prompts | No | No | Browser driver | No |
| Infrastructure to maintain | None, hosted | Grid | Yours | Yours | Yours |
Framework capabilities reflect what ships in the box. Nearly all of them can be extended with community libraries, which is exactly the maintenance we are offering to take off your plate.
Postman is a fine API client. The difference is whether the API check can live in the same scenario as the UI check.
| Feature | ReTibe | Postman | Frameworks |
|---|---|---|---|
| REST API testing | Yes | Yes | Through code |
| GraphQL testing | Query, mutation, subscription | Yes | Through code |
| JSON schema validation | Built-in | Scripts | Libraries |
| Rate limiting testing | Built-in | Manual | Manual |
| Auth: Bearer, Basic, API key | Built-in | Yes | Through code |
| Header validation with regex | Yes | Scripts | Through code |
| Variables from previous steps | {{variable}} |
Scripts | Through code |
| API plus UI in one scenario | Yes | No | Manual glue |
| Cookies reused from login | Automatic | Manual | Through code |
| Run history and analytics | Timeline, flakiness, trends | Paid cloud | No |
| Web interface, no code | JSON + visual | GUI | Code |
A single ReTibe scenario opens the page, calls the API, feeds the response into a form with {{variable}}, checks the result through the UI and produces one report. With frameworks and Postman that is two or three tools and manual stitching.
No framework ships this. You would reach for a separate scanner, on a separate schedule, owned by a separate person.
What you get back: a report scored 0 to 100, with a fix for each finding and a reference to the OWASP guidance it comes from. Eleven of the seventeen run by default; name the rest in the step to switch them on.
Schedules live in the product, not in someone's CI config. Colour-coded history and email alerts included.
Timeline, duration trend, stability heatmap per step and a 0 to 100 flakiness score, out of the box. Over MCP, two runs diff step by step, and each step's history tells a flaky step from one that regressed.
Pixel-perfect diff between the live page and the exported frame, so the build stays true to the design.
A bug created automatically when a test fails, with the report link and screenshots attached.
A green test that verifies nothing is worse than no test. Ours gets flagged in milliseconds, before a browser opens, along with data fields the engine never reads, selectors the browser cannot parse and {{params.*}} keys that do not exist. None of it spends a run.
Describe what to check and get a working scenario. Or send a request and the AI writes one for your site within two days.
When you need arbitrary code in the middle of a test, a custom runner, or logic no JSON schema can express. Frameworks are unbounded, and that is their real advantage. If your suite lives on that freedom, keep it.
No. Scenarios are plain JSON. Record them in the browser, describe them to the assistant, or let the agent generate them from your codebase. Selenium and Cypress need JavaScript or TypeScript and a build setup.
Yes. A CLI and a REST API plug into GitHub Actions, GitLab CI, Jenkins and Azure DevOps, and reports export as JUnit XML so your existing CI reads them unchanged. For GitHub Actions there is a ready action that runs one stored scenario on the platform and fails the job on its verdict, with no browser in your runner.
Yes. The free plan runs 20 web tests and 20 bot tests a month, with 5 templates, 2 security scans, 60 page inspections a month and 100 API requests a day. Enough for a small project and for judging whether it fits yours.
You can. It means learning Postman, Cypress and a separate scanner, then stitching three sets of results together by hand. ReTibe puts the checks in one scenario and one report.
Keep the framework you have. Take the one flow you would hate to see break, put it here, and read the report in the morning.