> For the complete documentation index, see [llms.txt](https://docs.drizz.dev/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.drizz.dev/reference/known-limitations.md).

# Known Limitations

Every documented limitation, grouped by where it applies. This page is the single source of truth for limits.

|                   |                                                                               |
| ----------------- | ----------------------------------------------------------------------------- |
| **Platform**      | `PRESS_DEVICE_BACK_BUTTON` doesn't exist on iOS; `CLEAR_APP` is a no-op there |
| **Language**      | No loops; a blocker rule carries one action; datasets nest two levels         |
| **Accessibility** | Visual checks only — no WCAG, screen reader, contrast or ARIA validation      |
| **Automation**    | No run-status API, no JUnit/JSON export, no webhooks                          |
| **Watch out**     | Secure screens (payment PIN entry) render black and can't be automated at all |

## Platform

| Limitation                                                                       | What to do instead                                         |
| -------------------------------------------------------------------------------- | ---------------------------------------------------------- |
| `PRESS_DEVICE_BACK_BUTTON` doesn't exist on iOS                                  | Use `Tap on the back button`, or a left-edge swipe         |
| `CLEAR_APP` is effectively a no-op on iOS                                        | Reinstall the build, or reset state in the app itself      |
| Pinch and zoom gestures are limited to specific Android emulator types — not iOS | Avoid pinch-dependent assertions in cross-platform tests   |
| iOS steps run roughly **twice as slow** as the equivalent Android steps          | Budget wall-clock time accordingly; keep iOS plans smaller |

## Language

| Limitation                                   | What to do instead                                             |
| -------------------------------------------- | -------------------------------------------------------------- |
| **No loops.** There's no iteration construct | Repeat the steps, or use a data-driven test plan               |
| A **blocker rule carries one action only**   | Use one rule per action, or handle the flow with an `IF` block |
| **Dataset nesting is limited to two levels** | Flatten the structure before you bind it                       |

## Chat

| Limitation                                                            | What to do instead                      |
| --------------------------------------------------------------------- | --------------------------------------- |
| **Android only** — Chat doesn't generate tests against iOS            | Author iOS tests by hand                |
| **Roughly 20–25 steps** per generated test                            | Split longer flows, or hand-author them |
| **No API steps, module calls, `Store` or `SET`** in generated scripts | Add these by hand after generation      |

Chat is also referred to as Fathom in older material — same feature.

## Accessibility

Drizz validates what is visually rendered on screen. Everything below that boundary is out of scope.

| Not supported                          | Detail                                                                                             |
| -------------------------------------- | -------------------------------------------------------------------------------------------------- |
| **WCAG conformance scanning**          | No compliance report. Nothing Drizz produces is evidence of legal conformance                      |
| **Screen reader testing**              | No TalkBack, VoiceOver or NVDA. Drizz does not drive assistive technology                          |
| **Contrast-ratio and color analysis**  | No contrast checks, color-blindness simulation or palette validation                               |
| **Semantic and role-level validation** | No inspection of ARIA roles, accessibility labels, alt text, focus order or the accessibility tree |

A dedicated accessibility tool is required for any of the above.

## Automation

| Missing                                        | Consequence                                                          |
| ---------------------------------------------- | -------------------------------------------------------------------- |
| **No API to poll run status or fetch results** | A pipeline can start a run but can't wait for it                     |
| **No JUnit, JSON or Allure export**            | Results can't be ingested by a CI reporter or a test-management tool |
| **No webhooks**                                | Nothing calls you back when a run finishes                           |

CI integration is fire-and-forget: a pipeline can trigger Drizz and cannot gate a build on the outcome. See [CI/CD](/automate-and-integrate/ci-cd.md).

## Environment

| Limitation                                                 | Detail                                                                                                                                                         |
| ---------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Secure screens can't be automated**                      | Payment PIN entry and some banking flows render black to any capture tool, Drizz included. There is no workaround — the screen contents don't exist to capture |
| **Location can be set at provisioning, not mid-test**      | Set it when the cloud device is provisioned. You can't change it partway through a run                                                                         |
| **Precise-location permission isn't grantable everywhere** | Availability varies by device target                                                                                                                           |

## Next

* [Changelog](/reference/changelog.md) — limitations lifted, newest first
* [Glossary](/reference/glossary.md) — terms used above
* [Troubleshooting](/reports-and-debugging/troubleshooting.md) — symptom, then fix

***

*Last updated: 6 August 2026*