> ## Documentation Index
> Fetch the complete documentation index at: https://docs.stably.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# No-code Tests vs Stably SDK

> Choose between Stably's no-code editor and Stably SDK to balance speed, scale, and customization.

Stably gives you two complementary ways to build Playwright-compatible tests. The no-code AI editor lets you record or generate tests in minutes and run them on Stably's managed cloud. Stably SDK drops into your Playwright repo when you need code-level control, advanced customization, and deeper CI/CD integration. Both share the same AI superpowers—AI assertions, AI actions, auto-heal, and smart selectors—so you can move between them without losing coverage.

## Side-by-side at a glance

| Criteria          | No-code AI Tests                                    | Stably SDK                                                |
| ----------------- | --------------------------------------------------- | --------------------------------------------------------- |
| **Primary focus** | Fastest path to reliable tests without code         | Full Playwright API plus Stably's AI layer                |
| **Setup effort**  | None—everything runs in Stably's cloud              | Install NPM package and manage Playwright config          |
| **How you build** | Record flows, generate with AI, drag-and-drop steps | Code in your IDE with Playwright fixtures and helpers     |
| **Collaboration** | Great for QA, PMs, and engineers working together   | Ideal for engineering-led teams building reusable modules |
| **Where it runs** | Managed cloud runners with zero infra               | Your CI/CD, local machines, or Stably cloud               |
| **Best for**      | Quick coverage, exploratory or UI workflows         | Complex apps, custom data setup, deep integrations        |

## When to stay in the no-code editor

* Launch your first suite quickly with recording and AI-generated steps—no repository or tooling required.
* Rely on Stably's cloud scale to run tests in parallel without maintaining infrastructure.
* Use AI assertions, AI actions, and auto-heal to express intent at a high level and reduce flakiness.
* Collaborate with teammates who prefer a visual editor while still exporting results, logs, and screenshots.
* Iterate on flows that change often; the editor's abstraction and healing keep maintenance low.

See the full workflow in [No-code AI Tests](/getting-started/no-code-tests).

## When to adopt Stably SDK

* You need the entire Playwright API along with custom fixtures, helpers, or page objects.
* Tests must live alongside application code with version control, code review, and branching.
* Your CI/CD pipeline, local debugging, or advanced test data orchestration demands code.
* You want to reuse existing Playwright suites while adding AI assertions, AI actions, and auto-heal.
* Teams require granular control over dependencies, secrets, and runtime environments.

Learn more in [Stably Playwright SDK](/getting-started/stably-playwright-sdk).

## Transition without rewrites

1. Build or prototype flows in the no-code editor to move fast and validate coverage.
2. Export any no-code test to Playwright when you need code-level control.
3. Replace `@playwright/test` with `@stablyai/playwright-test` to keep AI features in code.
4. Continue running tests in Stably's cloud or in your own infrastructure—the runners are interchangeable.

Because both experiences share the same execution engine, you can mix approaches in the same organization. Many teams keep high-level regression coverage in no-code while powering complex setup or API-heavy scenarios with the SDK.

## Recommended starting point

* Start in no-code for fast wins, onboarding non-developers, and validating core user journeys.
* Graduate individual tests—or entire suites—to Stably SDK when customization, reuse, or deep integrations outweigh the simplicity of the editor.
* Revisit no-code for exploratory coverage or when product teams want to contribute tests without writing code.