assuremind 1.3.0 → 1.4.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/README.md CHANGED
@@ -1,389 +1,452 @@
1
- # Assuremind Studio
2
-
3
- **AI-powered codeless UI & API test automation framework**
4
-
5
- [![npm version](https://img.shields.io/npm/v/assuremind.svg)](https://www.npmjs.com/package/assuremind)
6
- [![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)
7
- [![Node.js](https://img.shields.io/badge/node-%3E%3D18-brightgreen.svg)](https://nodejs.org)
8
- [![Playwright](https://img.shields.io/badge/powered%20by-Playwright-2EAD33.svg)](https://playwright.dev)
9
-
10
- Describe tests in plain English. AI generates Playwright code. Run anywhere.
11
-
12
- ---
13
-
14
- ## Why Assuremind?
15
-
16
- | Capability | What it does |
17
- |-----------|-------------|
18
- | **Zero coding** | Write steps in plain English — AI generates Playwright code automatically |
19
- | **MCP-sighted generation** | AI sees real page elements via Playwright MCP accessibility snapshots (~90-95% accuracy) |
20
- | **3 suite types** | **UI** (browser automation) · **API** (HTTP tests) · **Audit** (Playwright + Lighthouse) |
21
- | **5-level self-healing** | Broken selectors are auto-fixed by AI during runs — smart retry → AI regen → multi-selector → visual → decompose |
22
- | **12 AI providers** | Anthropic · OpenAI · Gemini · Groq · DeepSeek · Together · Qwen · Perplexity · Ollama · Bedrock · Azure OpenAI · Custom |
23
- | **Device emulation** | iPhone, Pixel, iPad, Galaxyfull Playwright device descriptors from UI or `--device` CLI flag |
24
- | **Studio UI** | Browser-based editor, run dashboard, reports, healing review, git control center with dark mode |
25
- | **RAG memory** | AI learns from every run retrieves similar past steps & healing fixes for smarter generation |
26
- | **Test Recorder** | Record tests by clicking in a real browserlocators verified against Playwright's accessibility tree, zero AI cost. Recorder + AI both handle iframes (click & fill), Shadow DOM, JS alerts, and keyboard actions |
27
- | **Cost-optimised** | Template engine + code cache + RAG handle ~80% of steps with zero AI calls |
28
- | **CI-ready** | `npx assuremind run --all --ci` exit code 0/1, works with any pipeline |
29
- | **File-based** | Plain JSON storage, fully Git-friendly, no database |
30
- | **Faker Data** | 100+ @faker-js/faker generators across 16 categories random emails, names, addresses, sequences no code, no stale data |
31
- | **Visual regression** | Pixel-diff screenshot comparison with baselinesapprove/reject visual changes in Studio |
32
- | **CI/CD integration** | Quality gates, PR comments (GitHub/GitLab), Slack/Teams notifications |
33
-
34
- ---
35
-
36
- ## Quick Start
37
-
38
- ```bash
39
- npm install assuremind
40
- npx assuremind init # folders, config, Playwright browsers
41
- npx assuremind studio # opens http://localhost:4400
42
- ```
43
-
44
- ### System Requirements
45
-
46
- | | Minimum | Recommended |
47
- |---|---|---|
48
- | **Node.js** | 18 LTS | 20 LTS or newer |
49
- | **OS** | macOS / Linux / Windows 10+ | — |
50
- | **RAM** | 4 GB | 8 GB+ |
51
- | **Disk** | ~2 GB (app + Playwright browsers) | 5 GB+ |
52
- | **Java** | — | JDK 17+ — only for Allure HTML reports ([adoptium.net](https://adoptium.net)) |
53
-
54
- `npx assuremind init` installs the Playwright browsers. Using a **cloud AI provider** (Anthropic, OpenAI, Gemini, …) needs only an API key and the specs above. Running a **local model via Ollama** has heavier requirements — see below.
55
-
56
- ### Configure AI Provider
57
-
58
- Edit `.env` pick one provider:
59
-
60
- ```bash
61
- # Anthropic # OpenAI # Google
62
- AI_PROVIDER=anthropic AI_PROVIDER=openai AI_PROVIDER=google
63
- ANTHROPIC_API_KEY=sk-ant-... OPENAI_API_KEY=sk-... GOOGLE_API_KEY=AIza...
64
- ANTHROPIC_MODEL=claude-sonnet.. OPENAI_MODEL=gpt-4o GOOGLE_MODEL=gemini-2.5-pro
65
- ```
66
-
67
- See `.env.example` for all 12 providers including Gemini, OpenAI, Anthropic, Ollama (local/free), etc.
68
-
69
- ### Local & Free — Ollama
70
-
71
- Run models entirely on your own machine (no API key, no cost). AssureMind uses the model to turn
72
- plain-English steps into Playwright code, so a **code-tuned model with strong instruction-following**
73
- works best.
74
-
75
- ```bash
76
- # 1. Install Ollama https://ollama.com/download
77
- ollama serve # starts the server on http://localhost:11434
78
- ollama pull qwen2.5-coder:7b # 2. pull the recommended model
79
- ```
80
-
81
- ```bash
82
- # 3. .env
83
- AI_PROVIDER=ollama
84
- OLLAMA_MODEL=qwen2.5-coder:7b
85
- OLLAMA_BASE_URL=http://localhost:11434
86
- AI_TIMEOUT=120 # local inference is slower than cloud APIs
87
- ```
88
-
89
- **Recommended models (by hardware):**
90
-
91
- | Hardware | Model | Notes |
92
- |----------|-------|-------|
93
- | 16 GB+ RAM, GPU 8 GB+ VRAM | `qwen2.5-coder:14b` | Highest accuracy — best for complex steps / iframes |
94
- | 16 GB RAM (GPU optional) ⭐ | `qwen2.5-coder:7b` | **Best balance** of quality & speed for this task |
95
- | 8 GB RAM, no GPU | `llama3.2` (3B) / `qwen2.5-coder:3b` | Usable; good for simpler steps |
96
- | ≤4–6 GB RAM | `llama3.2:1b` / `phi3:mini` | Lightweight; navigation & clicks only |
97
-
98
- **Ollama system requirements:**
99
-
100
- - **RAM is the floor** a model needs roughly its file size **+ ≈2 GB free**. A 7B model (≈4.7 GB)
101
- wants **8 GB free RAM minimum, 16 GB comfortable**; a 14B model wants **≈16 GB+**.
102
- - **GPU is optional but transformative**CPU-only works (hence `AI_TIMEOUT=120`), but a GPU with
103
- **8 GB+ VRAM** (NVIDIA CUDA / Apple Silicon / modern AMD) makes 7B models fast. Apple Silicon (M-series) is excellent thanks to unified memory.
104
- - **Disk** a few GB per model; keep **10–20 GB free** if trying several.
105
- - **Tips** enable **MCP** (Settings live page snapshots) for far more accurate locators with local
106
- models. Avoid `deepseek-r1` here: its `<think>` reasoning output is slow and pollutes generated code.
107
-
108
- ---
109
-
110
- ## CLI
111
-
112
- ```bash
113
- npx assuremind run --all # run everything
114
- npx assuremind run --type ui --tag smoke # filter by type + tag
115
- npx assuremind run --suite "Login" --browser chromium # run a suite
116
- npx assuremind run --all --device "iPhone 15 Pro" --ci # mobile + CI mode
117
- npx assuremind generate --story "User resets password" # AI generates full suite
118
- npx assuremind apply-healing --yes # accept all healed selectors
119
- npx assuremind validate # check config health
120
- npx assuremind doctor # system diagnostics
121
- ```
122
-
123
- | Flag | Description |
124
- |------|-------------|
125
- | `--all` | Run every suite |
126
- | `--type <ui\|api\|audit>` | Filter by suite type |
127
- | `--suite <name>` | Partial name match |
128
- | `--tag <tag>` | Filter by tag |
129
- | `--device <name>` | Emulate device (e.g. `"iPhone 15 Pro"`, `"Pixel 7"`) |
130
- | `--browser <list>` | `chromium` `firefox` `webkit` |
131
- | `--ci` | CI mode — exit code reflects pass/fail |
132
- | `--headed` | Show browser window |
133
- | `--no-healing` | Disable self-healing |
134
-
135
- Full reference [docs/CLI-REFERENCE.md](docs/CLI-REFERENCE.md)
136
-
137
- ---
138
-
139
- ## Studio UI
140
-
141
- Start with `npx assuremind studio` opens at `http://localhost:4400`.
142
-
143
- **Dashboard** · **Smart Tests** · **Test Editor** · **Run Config** · **Reports** · **Variables** · **Self-Healing** · **Step Library** · **Faker Data** · **CI/CD** · **Git Control** · **Settings** · **Docs**
144
-
145
- Full walkthrough → [docs/STUDIO.md](docs/STUDIO.md)
146
-
147
- ---
148
-
149
- ## MCP Integration
150
-
151
- AI sees **real page elements** during code generation via the official `@playwright/mcp` server. Enabled by default.
152
-
153
- | Mode | Accuracy | Latency | Config |
154
- |------|----------|---------|--------|
155
- | Blind (MCP off) | ~50-70% | Fastest | `mcp.enabled: false` |
156
- | **Snapshot-driven** | ~90-95% | +2-5s first page | `mcp.enabled: true` (default) |
157
- | Act-then-script | ~98-100% | +5-10s/step | `mcp.actThenScript: true` |
158
-
159
- - MCP is **only used during code generation** — test execution is never affected
160
- - Silent fallback if MCP fails, generation continues blindly without error
161
- - Configure in **Settings → MCP Integration** or `autotest.config.json`
162
-
163
- ---
164
-
165
- ## Test Recorder
166
-
167
- Record tests by interacting with your application in a real browser — **zero AI, zero cost, zero guesswork**.
168
-
169
- Click **Record** in the Test Editor, perform your actions, and stop. Each interaction becomes a step with verified Playwright code, ready to run.
170
-
171
- ### How it works
172
-
173
- > Available for **UI** and **Audit** suites only (not API suites).
174
-
175
- 1. A headed Chromium browser opens your app's URL
176
- 2. Every click, fill, navigation, and keyboard action is captured in real time — **including inside iframes and shadow roots**
177
- 3. Locators are resolved against Playwright's **accessibility tree** — the recorder tries 6 strategies (data-testid, getByRole, getByLabel, getByPlaceholder, getByText, CSS) and verifies each with `count() === 1`
178
- 4. **Iframe-aware** elements inside iframes automatically generate `page.frameLocator('#iframe').getByRole(...)` code with the correct frame chain
179
- 5. **Shadow DOM** — `composedPath()[0]` pierces shadow roots; generates `host >> inner` pierce locators automatically
180
- 6. **JS dialogs** — alerts/confirms are auto-accepted during recording; `page.once('dialog', ...)` is prepended to the triggering step's code
181
- 7. **Keyboard** — Tab, Shift+Tab, arrows, Enter, Escape, Ctrl+A, and Space on buttons are all captured
182
- 8. Assertions via keyboard shortcuts: **Shift+Click** (element visible), **Ctrl+Shift+U** (URL), **Ctrl+Shift+T** (page title)
183
- 9. On stop, each action is added as a step with **pre-generated Playwright code** — no AI call needed
184
-
185
- ### What makes it stand out vs other recorders
186
-
187
- | Feature | Selenium IDE | Playwright Codegen | Assuremind Recorder |
188
- |---------|-------------|-------------------|---------------------|
189
- | Locator quality | CSS/XPath | Good | Best 6 strategies, verified against live page |
190
- | Accessibility tree | No | Partial | Full every locator checked via Playwright API |
191
- | **Iframe support** | Partial | Manual | **Full** detects iframes, generates `frameLocator()` code — click **and** fill both work |
192
- | **Shadow DOM** | No | Limited | `>>` pierce combinator elements inside shadow roots are reachable |
193
- | **JS Alert handling** | No | No | `page.once('dialog', ...)` registered automatically before triggering action |
194
- | **Keyboard actions** | No | No | `page.keyboard.press()` / `.type()` for keys, shortcuts, and character input |
195
- | Assertions | Manual | Manual | Shift+Click (hard), Ctrl+Shift+Click (soft), URL & title shortcuts |
196
- | Plain-English steps | No | No | Yes human-readable instructions auto-generated |
197
- | Self-healing after | No | No | Yes — 5-level AI healing cascade |
198
- | RAG memory | No | No | Yesrecorded steps feed the learning loop |
199
- | Cost | Free | Free | Free |
200
-
201
- ### Biggest pain points in test automationsolved
202
-
203
- | Pain Point | How the Recorder Solves It |
204
- |-----------|---------------------------|
205
- | Writing tests is slow | Record a full test in 30 seconds |
206
- | Selectors break constantly | Locators verified against Playwright's accessibility tree in real time |
207
- | AI costs money | Recording + code generation = $0, zero AI calls |
208
- | Non-technical testers can't write tests | Anyone who can click a browser can create tests |
209
- | Assertions are hard to write | Shift+Click for hard, Ctrl+Shift+Click for soft, Ctrl+Shift+U for URL, Ctrl+Shift+T for title |
210
- | Hard vs soft assertions | Soft assertions (`expect.soft()`) let the test continueall failures reported at end |
211
- | Recorded tests are fragile | 6-strategy locator resolution + post-run 5-level self-healing |
212
- | Apps use iframes (SAP, Salesforce) | Recorder auto-detects iframe context, generates `frameLocator()` chains — **click and fill** both work |
213
- | Shadow DOM / Web Components | Recorder pierces shadow roots via `composedPath()`, generates `>>` locators automatically |
214
- | JavaScript alerts & popups | Recorder auto-accepts dialogs; `page.once('dialog', ...)` prepended to triggering step |
215
- | Keyboard interactions | Recorder captures Tab, Shift+Tab, arrows, Ctrl+A, Enter, Escape, and Space |
216
-
217
- ---
218
-
219
- ## RAG Memory (Retrieval-Augmented Generation)
220
-
221
- The AI learns from every test run, building semantic memory that improves accuracy over time — **zero setup required**:
222
-
223
- | Corpus | What it stores | When it's used |
224
- |--------|---------------|----------------|
225
- | **Code Corpus** | Instruction-to-code mappings from successful runs | During generation — similar past steps are retrieved as AI examples or used directly (score >= 0.90) |
226
- | **Healing Corpus** | Past healing events (error + fix pairs) | During self-healing — proven past fixes are injected into the repair prompt |
227
- | **Error Catalog** | Recurring error patterns per URL | During generation — the AI is warned about known-bad selectors to avoid |
228
-
229
- **Zero cost, zero database** — uses local TF-IDF embeddings and file-based JSON storage (`results/.rag/`). Enabled by default — works automatically from the very first run.
230
-
231
- ### How it improves over time
232
-
233
- - **Run 1** — memory is empty, AI generates code normally
234
- - **Run 2+** RAG kicks in silently: similar instructions are retrieved instead of making API calls (free + faster), healing uses proven past fixes
235
- - **Run 10+** most common steps are served from RAG memory at zero cost, self-healing resolves issues on the first attempt
236
-
237
- ### Consumer FAQ
238
-
239
- | Question | Answer |
240
- |----------|--------|
241
- | Do I need to configure anything? | No. RAG is ON by default with zero setup. |
242
- | Does it cost anything? | No. TF-IDF embedder runs locally. RAG direct hits replace paid AI calls. |
243
- | Does it slow down my tests? | No. RAG lookup is <1ms. It actually speeds up generation. |
244
- | Does it work in CI/CD? | Yes. Cache `results/.rag/` between CI runs to persist memory. |
245
- | How do I share memory across team? | Commit `results/.rag/` to Git or use a CI cache step. |
246
- | How do I reset memory? | Delete the `results/.rag/` folder. |
247
-
248
- ### When to use Settings → RAG Memory
249
-
250
- Most users never need to touch RAG settings. The Settings card exists for power-user scenarios:
251
-
252
- | Scenario | Action |
253
- |----------|--------|
254
- | Debugging a flaky test | Turn OFF Code Corpus forces fresh AI generation |
255
- | Healing keeps suggesting a bad fix | Turn OFF Healing Corpus — clears bad fix influence |
256
- | Major app redesign | Turn OFF RAG entirely — old memory is now misleading |
257
- | Error warnings are outdated | Turn OFF Error Catalog — stops avoiding selectors that are fine now |
258
- | Want deterministic CI runs | Disable RAG in CI config, keep ON locally |
259
-
260
- ---
261
-
262
- ## Self-Healing
263
-
264
- When your app's UI changes (button renamed, element moved, DOM restructured), tests break. Instead of failing immediately, Assuremind automatically attempts to fix the broken selector through a 5-level cascade **fully automated, no manual intervention needed**:
265
-
266
- | Level | What happens | Example | AI Cost |
267
- |-------|-------------|---------|---------|
268
- | 1 | **Smart retry** — waits for the element with exponential backoff | Element was loading slowly; retry finds it after 2s | Free |
269
- | 2 | **AI regeneration** — AI rewrites the Playwright code using current page context | `#login-btn` removed → AI generates `page.getByRole('button', { name: 'Sign In' })` | 1 call |
270
- | 3 | **Multi-selector** — AI generates 5 alternative selector strategies, tries each | Tries role → label → placeholder → text → CSS until one works | 1 call |
271
- | 4 | **Visual analysis** — takes a screenshot, AI visually locates the element | Button has no text/role but AI sees it in the screenshot | 1 call |
272
- | 5 | **Decompose** — breaks the failing step into 3-5 simpler micro-actions | "Fill login form and submit" → separate fill email + fill password + click submit | 1 call |
273
-
274
- If all 5 levels fail, the step is marked **failed** and saved to the healing report for your review.
275
-
276
- ### How you use it
277
-
278
- 1. **During test runs** — healing happens automatically. If Level 2 fixes a broken `#login-btn`, your test **passes** and continues.
279
- 2. **After the run** — healed selectors are saved as **pending suggestions** (not auto-applied to source files).
280
- 3. **Review & accept** — in Studio **Self-Healing** page, or from CLI:
281
- ```bash
282
- npx assuremind apply-healing # interactive review: accept/reject each fix
283
- npx assuremind apply-healing --yes # accept all in CI
284
- ```
285
- 4. **Accepted fixes** are written back to your `.test.json` files — next run uses the healed code permanently.
286
-
287
- > **CI/CD tip:** Add `npx assuremind apply-healing --yes` as a post-test step so healed selectors are committed back automatically. Enable `healing.autoPR` in Settings to auto-create a GitHub PR with the fixes.
288
-
289
- ---
290
-
291
- ## CI/CD
292
-
293
- ```yaml
294
- # GitHub Actions
295
- - name: Run tests
296
- env:
297
- AI_PROVIDER: google
298
- GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
299
- GOOGLE_MODEL: gemini-2.5-pro
300
- run: npx assuremind run --all --ci
301
- ```
302
-
303
- Also supports **GitLab CI** and **Jenkins** — or use the built-in **CI Config Generator** in Studio (Run Config → Generate CI Config).
304
-
305
- Exit code `0` = all passed · `1` = failures.
306
-
307
- ---
308
-
309
- ## Enterprise Features
310
-
311
- ### Faker Data
312
-
313
- 100+ @faker-js/faker generators across 16 categories — generate realistic random data without code.
314
-
315
- | Generator | Example output |
316
- |-----------|---------------|
317
- | `person.fullName` | Sarah Johnson |
318
- | `internet.email` | sarah.johnson@gmail.com |
319
- | `phone.number` | +1-555-123-4567 |
320
- | `finance.amount` | 249.99 |
321
- | `location.city` | San Francisco |
322
- | Sequence `USER-{n}` | USER-001, USER-002, USER-003 |
323
-
324
- Browse by category or search, select multiple tokens at once — selected tokens appear as removable chips, and clicking "Insert N tokens" inserts them all comma-separated (e.g. `{{FAKE_PERSON_FIRSTNAME}}, {{FAKE_PERSON_LASTNAME}}, {{FAKE_INTERNET_EMAIL}}`). Great for form filling. Tokens resolve to fresh random values on every test run — no more "email already exists" failures.
325
-
326
- ### Visual Regression
327
-
328
- Pixel-perfect screenshot comparison catch UI changes that functional tests miss. Configure in **Settings** → Visual Regression. Review diffs in **Reports** → Visual Diffs tab.
329
-
330
- 1. Enable in **Settings** Visual Regression, then toggle the **Eye icon** on any step in the Test Editor
331
- 2. First run captures **baseline** screenshots (committed to git)
332
- 3. Subsequent runs **compare** pixel-by-pixel using `pixelmatch`
333
- 4. Diff > threshold → step **fails** with highlighted diff image
334
- 5. Review in **Reports** → Visual Diffs → **Approve** (update baseline) or **Reject** (it's a bug)
335
-
336
- ```
337
- baselines/ # committed to git
338
- ├── login-suite/login-test/
339
- │ ├── step-3-chromium-1920x1080.png
340
- │ └── step-3-firefox-1920x1080.png
341
- results/visual-diffs/{runId}/ # gitignored, per-run
342
- ├── step-3-baseline.png
343
- ├── step-3-actual.png
344
- └── step-3-diff.png # red highlights
345
- ```
346
-
347
- ### CI/CD Integration
348
-
349
- Quality gates, PR comments, and notifications — configured from Studio, executed in pipelines.
350
-
351
- | Feature | What it does |
352
- |---------|-------------|
353
- | **Quality gates** | Set min pass rate (95%), max duration, required tags — pipeline fails if gate fails |
354
- | **PR comments** | Auto-post test results as GitHub PR comment or GitLab MR note |
355
- | **Notifications** | Slack, Teams, Email, custom webhook — trigger on failure, every run, or healing events |
356
-
357
- ```yaml
358
- # GitHub Actions — full integration
359
- - name: Run tests
360
- env:
361
- AI_PROVIDER: anthropic
362
- ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
363
- run: npx assuremind run --all --ci
364
- # Quality gate auto-evaluates, PR comment auto-posts, Slack notifies on failure
365
- ```
366
-
367
- ---
368
-
369
- ## Documentation
370
-
371
- | Resource | Location |
372
- |----------|----------|
373
- | Getting started | [docs/GETTING-STARTED.md](docs/GETTING-STARTED.md) |
374
- | Studio walkthrough | [docs/STUDIO.md](docs/STUDIO.md) |
375
- | CLI reference | [docs/CLI-REFERENCE.md](docs/CLI-REFERENCE.md) |
376
- | Contributing | [CONTRIBUTING.md](CONTRIBUTING.md) |
377
- | All AI providers | `.env.example` |
378
- | Built-in docs | Studio → **Docs** page |
379
- | Enterprise features | Studio **Faker Data**, **CI/CD** pages; Visual Regression in **Settings** + **Reports** |
380
-
381
- ---
382
-
383
- ## License
384
-
385
- MIT — see [LICENSE](LICENSE) for details.
386
-
387
- ---
388
-
389
- *Built by [Deepak Hiremath](https://www.linkedin.com/in/deepak-hiremath-0017937a/)*
1
+ # AssureMind Studio
2
+
3
+ **AI-powered codeless UI & API test automation framework**
4
+
5
+ [![npm version](https://img.shields.io/npm/v/assuremind.svg)](https://www.npmjs.com/package/assuremind)
6
+ [![License: Apache 2.0](https://img.shields.io/badge/License-Apache_2.0-blue.svg)](LICENSE)
7
+ [![Node.js](https://img.shields.io/badge/node-%3E%3D18-brightgreen.svg)](https://nodejs.org)
8
+ [![Playwright](https://img.shields.io/badge/powered%20by-Playwright-2EAD33.svg)](https://playwright.dev)
9
+
10
+ Describe tests in plain English. AI generates Playwright code. Run anywhere.
11
+
12
+ > ### 🌐 Website — [https://assuremind.in/](https://assuremind.in/)
13
+ > ### 💬 Community Hub — [https://assuremind.in/community/](https://assuremind.in/community/)
14
+ > Report bugs · request features · share ideas & feedback · ask questions.
15
+ > 📧 Contact: [assuremind.contact@gmail.com](mailto:assuremind.contact@gmail.com)
16
+
17
+ ---
18
+
19
+ ## Why AssureMind?
20
+
21
+ | Capability | What it does |
22
+ |-----------|-------------|
23
+ | **Zero coding** | Write steps in plain English AI generates Playwright code automatically |
24
+ | **MCP-sighted generation** | AI sees real page elements via Playwright MCP accessibility snapshots (~90-95% accuracy) |
25
+ | **3 suite types** | **UI** (browser automation) · **API** (HTTP tests) · **Audit** (Playwright + Lighthouse) |
26
+ | **5-level self-healing** | Broken selectors are auto-fixed by AI during runssmart retry AI regen multi-selector visual decompose |
27
+ | **12 AI providers** | Anthropic · OpenAI · Gemini · Groq · DeepSeek · Together · Qwen · Perplexity · Ollama · Bedrock · Azure OpenAI · Custom |
28
+ | **Device emulation** | iPhone, Pixel, iPad, Galaxyfull Playwright device descriptors from UI or `--device` CLI flag |
29
+ | **Studio UI** | Browser-based editor, run dashboard, reports, healing review, git control center — with dark mode |
30
+ | **RAG memory** | AI learns from every runretrieves similar past steps & healing fixes for smarter generation |
31
+ | **Test Recorder** | Record tests by clicking in a real browser locators verified against Playwright's accessibility tree, zero AI cost. Recorder + AI both handle iframes (click & fill), Shadow DOM, JS alerts, and keyboard actions |
32
+ | **Cost-optimised** | Template engine + code cache + RAG handle ~80% of steps with zero AI calls |
33
+ | **CI-ready** | `npx assuremind run --all --ci` — exit code 0/1, works with any pipeline |
34
+ | **File-based** | Plain JSON storage, fully Git-friendly, no database |
35
+ | **Faker Data** | 100+ @faker-js/faker generators across 16 categories — random emails, names, addresses, sequences — no code, no stale data |
36
+ | **File upload & download** | Upload fixtures from the Studio and insert as `{{FILE:…}}` tokens; downloads auto-captured per run and viewable in Reports |
37
+ | **Visual regression** | Pixel-diff screenshot comparison with baselines — approve/reject visual changes in Studio |
38
+ | **CI/CD integration** | Quality gates, PR comments (GitHub/GitLab), Slack/Teams notifications |
39
+
40
+ ---
41
+
42
+ ## Quick Start
43
+
44
+ ```bash
45
+ npm install assuremind
46
+ npx assuremind init # folders, config, Playwright browsers
47
+ npx playwright install # if browser install was skipped or failed during init
48
+ npx assuremind studio # opens http://localhost:4400
49
+ ```
50
+
51
+ > **First-time setup:** `npx assuremind init` installs the Playwright browsers automatically. If that step is skipped (`--skip-playwright`) or fails (network/permissions), install them manually with `npx playwright install` (add `--with-deps` on Linux to pull OS libraries). Run `npx assuremind doctor` to verify your setup.
52
+
53
+ ### System Requirements
54
+
55
+ | | Minimum | Recommended |
56
+ |---|---|---|
57
+ | **Node.js** | 18 LTS | 20 LTS or newer |
58
+ | **OS** | macOS / Linux / Windows 10+ | — |
59
+ | **RAM** | 4 GB | 8 GB+ |
60
+ | **Disk** | ~2 GB (app + Playwright browsers) | 5 GB+ |
61
+ | **Java** | — | JDK 17+ — only for Allure HTML reports |
62
+
63
+ `npx assuremind init` installs the Playwright browsers. Using a **cloud AI provider** (Anthropic, OpenAI, Gemini, …) needs only an API key and the specs above. Running a **local model via Ollama** has heavier requirements — see below.
64
+
65
+ ### Configure AI Provider
66
+
67
+ Edit `.env` pick one provider:
68
+
69
+ ```bash
70
+ # Anthropic # OpenAI # Google
71
+ AI_PROVIDER=anthropic AI_PROVIDER=openai AI_PROVIDER=google
72
+ ANTHROPIC_API_KEY=sk-ant-... OPENAI_API_KEY=sk-... GOOGLE_API_KEY=AIza...
73
+ ANTHROPIC_MODEL=claude-sonnet.. OPENAI_MODEL=gpt-4o GOOGLE_MODEL=gemini-2.5-pro
74
+ ```
75
+
76
+ See `.env.example` for all 12 providers including Gemini, OpenAI, Anthropic, Ollama (local/free), etc.
77
+
78
+ ### Local & Free Ollama
79
+
80
+ Run models entirely on your own machine (no API key, no cost). AssureMind uses the model to turn
81
+ plain-English steps into Playwright code, so a **code-tuned model with strong instruction-following**
82
+ works best.
83
+
84
+ ```bash
85
+ # 1. Install Ollama → https://ollama.com/download
86
+ ollama serve # starts the server on http://localhost:11434
87
+ ollama pull qwen2.5-coder:7b # 2. pull the recommended model
88
+ ```
89
+
90
+ ```bash
91
+ # 3. .env
92
+ AI_PROVIDER=ollama
93
+ OLLAMA_MODEL=qwen2.5-coder:7b
94
+ OLLAMA_BASE_URL=http://localhost:11434
95
+ AI_TIMEOUT=120 # local inference is slower than cloud APIs
96
+ ```
97
+
98
+ **Recommended models (by hardware):**
99
+
100
+ | Hardware | Model | Notes |
101
+ |----------|-------|-------|
102
+ | 16 GB+ RAM, GPU 8 GB+ VRAM | `qwen2.5-coder:14b` | Highest accuracy best for complex steps / iframes |
103
+ | 16 GB RAM (GPU optional) | `qwen2.5-coder:7b` | **Best balance** of quality & speed for this task |
104
+ | 8 GB RAM, no GPU | `llama3.2` (3B) / `qwen2.5-coder:3b` | Usable; good for simpler steps |
105
+ | ≤4–6 GB RAM | `llama3.2:1b` / `phi3:mini` | Lightweight; navigation & clicks only |
106
+
107
+ **Ollama system requirements:**
108
+
109
+ - **RAM is the floor** — a model needs roughly its file size **+ ≈2 GB free**. A 7B model (≈4.7 GB)
110
+ wants **8 GB free RAM minimum, 16 GB comfortable**; a 14B model wants **≈16 GB+**.
111
+ - **GPU is optional but transformative** — CPU-only works (hence `AI_TIMEOUT=120`), but a GPU with
112
+ **8 GB+ VRAM** (NVIDIA CUDA / Apple Silicon / modern AMD) makes 7B models fast. Apple Silicon (M-series) is excellent thanks to unified memory.
113
+ - **Disk** a few GB per model; keep **10–20 GB free** if trying several.
114
+ - **Tips** enable **MCP** (Settings live page snapshots) for far more accurate locators with local
115
+ models. Avoid `deepseek-r1` here: its `<think>` reasoning output is slow and pollutes generated code.
116
+
117
+ ---
118
+
119
+ ## CLI
120
+
121
+ ```bash
122
+ npx assuremind run --all # run everything
123
+ npx assuremind run --type ui --tag smoke # filter by type + tag
124
+ npx assuremind run --suite "Login" --browser chromium # run a suite
125
+ npx assuremind run --all --device "iPhone 15 Pro" --ci # mobile + CI mode
126
+ npx assuremind generate --story "User resets password" # AI generates full suite
127
+ npx assuremind apply-healing --yes # accept all healed selectors
128
+ npx assuremind validate # check config health
129
+ npx assuremind doctor # system diagnostics
130
+ ```
131
+
132
+ | Flag | Description |
133
+ |------|-------------|
134
+ | `--all` | Run every suite |
135
+ | `--type <ui\|api\|audit>` | Filter by suite type |
136
+ | `--suite <name>` | Partial name match |
137
+ | `--tag <tag>` | Filter by tag |
138
+ | `--device <name>` | Emulate device (e.g. `"iPhone 15 Pro"`, `"Pixel 7"`) |
139
+ | `--browser <list>` | `chromium` `firefox` `webkit` |
140
+ | `--ci` | CI mode — exit code reflects pass/fail |
141
+ | `--headed` | Show browser window |
142
+ | `--no-healing` | Disable self-healing |
143
+
144
+ Full reference → [docs/CLI-REFERENCE.md](docs/CLI-REFERENCE.md)
145
+
146
+ ---
147
+
148
+ ## Studio UI
149
+
150
+ Start with `npx assuremind studio` — opens at `http://localhost:4400`.
151
+
152
+ **Dashboard** · **Smart Tests** · **Test Editor** · **Run Config** · **Reports** · **Variables** · **Self-Healing** · **Step Library** · **Faker Data** · **CI/CD** · **Git Control** · **Settings** · **Docs**
153
+
154
+ Full walkthrough → [docs/STUDIO.md](docs/STUDIO.md)
155
+
156
+ ---
157
+
158
+ ## MCP Integration
159
+
160
+ AI sees **real page elements** during code generation via the official `@playwright/mcp` server. Enabled by default.
161
+
162
+ | Mode | Accuracy | Latency | Config |
163
+ |------|----------|---------|--------|
164
+ | Blind (MCP off) | ~50-70% | Fastest | `mcp.enabled: false` |
165
+ | **Snapshot-driven** | ~90-95% | +2-5s first page | `mcp.enabled: true` (default) |
166
+ | Act-then-script | ~98-100% | +5-10s/step | `mcp.actThenScript: true` |
167
+
168
+ - MCP is **only used during code generation** — test execution is never affected
169
+ - Silent fallback if MCP fails, generation continues blindly without error
170
+ - Configure in **Settings → MCP Integration** or `autotest.config.json`
171
+
172
+ ---
173
+
174
+ ## Test Recorder
175
+
176
+ Record tests by interacting with your application in a real browser — **zero AI, zero cost, zero guesswork**.
177
+
178
+ Click **Record** in the Test Editor, perform your actions, and stop. Each interaction becomes a step with verified Playwright code, ready to run.
179
+
180
+ ### How it works
181
+
182
+ > Available for **UI** and **Audit** suites only (not API suites).
183
+
184
+ 1. A headed Chromium browser opens your app's URL
185
+ 2. Every click, fill, navigation, and keyboard action is captured in real time — **including inside iframes and shadow roots**
186
+ 3. Locators are resolved against Playwright's **accessibility tree** — the recorder tries 6 strategies (data-testid, getByRole, getByLabel, getByPlaceholder, getByText, CSS) and verifies each with `count() === 1`
187
+ 4. **Iframe-aware** elements inside iframes automatically generate `page.frameLocator('#iframe').getByRole(...)` code with the correct frame chain
188
+ 5. **Shadow DOM** — `composedPath()[0]` pierces shadow roots; generates `host >> inner` pierce locators automatically
189
+ 6. **JS dialogs** alerts/confirms are auto-accepted during recording; `page.once('dialog', ...)` is prepended to the triggering step's code
190
+ 7. **Keyboard** Tab, Shift+Tab, arrows, Enter, Escape, Ctrl+A, and Space on buttons are all captured
191
+ 8. Assertions via keyboard shortcuts: **Shift+Click** (element visible), **Ctrl+Shift+U** (URL), **Ctrl+Shift+T** (page title)
192
+ 9. On stop, each action is added as a step with **pre-generated Playwright code** no AI call needed
193
+
194
+ ### What makes it stand out vs other recorders
195
+
196
+ | Feature | Selenium IDE | Playwright Codegen | AssureMind Recorder |
197
+ |---------|-------------|-------------------|---------------------|
198
+ | Locator quality | CSS/XPath | Good | Best6 strategies, verified against live page |
199
+ | Accessibility tree | No | Partial | Full — every locator checked via Playwright API |
200
+ | **Iframe support** | Partial | Manual | **Full** — detects iframes, generates `frameLocator()` code — click **and** fill both work |
201
+ | **Shadow DOM** | No | Limited | `>>` pierce combinator elements inside shadow roots are reachable |
202
+ | **JS Alert handling** | No | No | `page.once('dialog', ...)` registered automatically before triggering action |
203
+ | **Keyboard actions** | No | No | `page.keyboard.press()` / `.type()` for keys, shortcuts, and character input |
204
+ | Assertions | Manual | Manual | Shift+Click (hard), Ctrl+Shift+Click (soft), URL & title shortcuts |
205
+ | Plain-English steps | No | No | Yes human-readable instructions auto-generated |
206
+ | Self-healing after | No | No | Yes 5-level AI healing cascade |
207
+ | RAG memory | No | No | Yes recorded steps feed the learning loop |
208
+ | Cost | Free | Free | Free |
209
+
210
+ ### Biggest pain points in test automationsolved
211
+
212
+ | Pain Point | How the Recorder Solves It |
213
+ |-----------|---------------------------|
214
+ | Writing tests is slow | Record a full test in 30 seconds |
215
+ | Selectors break constantly | Locators verified against Playwright's accessibility tree in real time |
216
+ | AI costs money | Recording + code generation = $0, zero AI calls |
217
+ | Non-technical testers can't write tests | Anyone who can click a browser can create tests |
218
+ | Assertions are hard to write | Shift+Click for hard, Ctrl+Shift+Click for soft, Ctrl+Shift+U for URL, Ctrl+Shift+T for title |
219
+ | Hard vs soft assertions | Soft assertions (`expect.soft()`) let the test continue — all failures reported at end |
220
+ | Recorded tests are fragile | 6-strategy locator resolution + post-run 5-level self-healing |
221
+ | Apps use iframes (SAP, Salesforce) | Recorder auto-detects iframe context, generates `frameLocator()` chains — **click and fill** both work |
222
+ | Shadow DOM / Web Components | Recorder pierces shadow roots via `composedPath()`, generates `>>` locators automatically |
223
+ | JavaScript alerts & popups | Recorder auto-accepts dialogs; `page.once('dialog', ...)` prepended to triggering step |
224
+ | Keyboard interactions | Recorder captures Tab, Shift+Tab, arrows, Ctrl+A, Enter, Escape, and Space |
225
+
226
+ ---
227
+
228
+ ## RAG Memory (Retrieval-Augmented Generation)
229
+
230
+ The AI learns from every test run, building semantic memory that improves accuracy over time — **zero setup required**:
231
+
232
+ | Corpus | What it stores | When it's used |
233
+ |--------|---------------|----------------|
234
+ | **Code Corpus** | Instruction-to-code mappings from successful runs | During generation — similar past steps are retrieved as AI examples or used directly (score >= 0.90) |
235
+ | **Healing Corpus** | Past healing events (error + fix pairs) | During self-healing proven past fixes are injected into the repair prompt |
236
+ | **Error Catalog** | Recurring error patterns per URL | During generation — the AI is warned about known-bad selectors to avoid |
237
+
238
+ **Zero cost, zero database** — uses local TF-IDF embeddings and file-based JSON storage (`results/.rag/`). Enabled by default — works automatically from the very first run.
239
+
240
+ ### How it improves over time
241
+
242
+ - **Run 1** memory is empty, AI generates code normally
243
+ - **Run 2+** RAG kicks in silently: similar instructions are retrieved instead of making API calls (free + faster), healing uses proven past fixes
244
+ - **Run 10+** most common steps are served from RAG memory at zero cost, self-healing resolves issues on the first attempt
245
+
246
+ ### Consumer FAQ
247
+
248
+ | Question | Answer |
249
+ |----------|--------|
250
+ | Do I need to configure anything? | No. RAG is ON by default with zero setup. |
251
+ | Does it cost anything? | No. TF-IDF embedder runs locally. RAG direct hits replace paid AI calls. |
252
+ | Does it slow down my tests? | No. RAG lookup is <1ms. It actually speeds up generation. |
253
+ | Does it work in CI/CD? | Yes. Cache `results/.rag/` between CI runs to persist memory. |
254
+ | How do I share memory across team? | Commit `results/.rag/` to Git or use a CI cache step. |
255
+ | How do I reset memory? | Delete the `results/.rag/` folder. |
256
+
257
+ ### When to use Settings RAG Memory
258
+
259
+ Most users never need to touch RAG settings. The Settings card exists for power-user scenarios:
260
+
261
+ | Scenario | Action |
262
+ |----------|--------|
263
+ | Debugging a flaky test | Turn OFF Code Corpus — forces fresh AI generation |
264
+ | Healing keeps suggesting a bad fix | Turn OFF Healing Corpusclears bad fix influence |
265
+ | Major app redesign | Turn OFF RAG entirely — old memory is now misleading |
266
+ | Error warnings are outdated | Turn OFF Error Catalog — stops avoiding selectors that are fine now |
267
+ | Want deterministic CI runs | Disable RAG in CI config, keep ON locally |
268
+
269
+ ---
270
+
271
+ ## Self-Healing
272
+
273
+ When your app's UI changes (button renamed, element moved, DOM restructured), tests break. Instead of failing immediately, AssureMind automatically attempts to fix the broken selector through a 5-level cascade — **fully automated, no manual intervention needed**:
274
+
275
+ | Level | What happens | Example | AI Cost |
276
+ |-------|-------------|---------|---------|
277
+ | 1 | **Smart retry** — waits for the element with exponential backoff | Element was loading slowly; retry finds it after 2s | Free |
278
+ | 2 | **AI regeneration** AI rewrites the Playwright code using current page context | `#login-btn` removed AI generates `page.getByRole('button', { name: 'Sign In' })` | 1 call |
279
+ | 3 | **Multi-selector**AI generates 5 alternative selector strategies, tries each | Tries role → label → placeholder → text → CSS until one works | 1 call |
280
+ | 4 | **Visual analysis** takes a screenshot, AI visually locates the element | Button has no text/role but AI sees it in the screenshot | 1 call |
281
+ | 5 | **Decompose** — breaks the failing step into 3-5 simpler micro-actions | "Fill login form and submit" → separate fill email + fill password + click submit | 1 call |
282
+
283
+ If all 5 levels fail, the step is marked **failed** and saved to the healing report for your review.
284
+
285
+ ### How you use it
286
+
287
+ 1. **During test runs** healing happens automatically. If Level 2 fixes a broken `#login-btn`, your test **passes** and continues.
288
+ 2. **After the run** — healed selectors are saved as **pending suggestions** (not auto-applied to source files).
289
+ 3. **Review & accept** — in Studio → **Self-Healing** page, or from CLI:
290
+ ```bash
291
+ npx assuremind apply-healing # interactive review: accept/reject each fix
292
+ npx assuremind apply-healing --yes # accept all in CI
293
+ ```
294
+ 4. **Accepted fixes** are written back to your `.test.json` files — next run uses the healed code permanently.
295
+
296
+ > **CI/CD tip:** Add `npx assuremind apply-healing --yes` as a post-test step so healed selectors are committed back automatically. Enable `healing.autoPR` in Settings to auto-create a GitHub PR with the fixes.
297
+
298
+ ---
299
+
300
+ ## CI/CD
301
+
302
+ ```yaml
303
+ # GitHub Actions
304
+ - name: Run tests
305
+ env:
306
+ AI_PROVIDER: google
307
+ GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
308
+ GOOGLE_MODEL: gemini-2.5-pro
309
+ run: npx assuremind run --all --ci
310
+ ```
311
+
312
+ Also supports **GitLab CI** and **Jenkins** — or use the built-in **CI Config Generator** in Studio (Run Config → Generate CI Config).
313
+
314
+ **CI Config Generator** (Run Config → Generate CI Config):
315
+ - Generates a workflow for **GitHub Actions**, **GitLab CI**, or **Jenkins** from your current run config.
316
+ - The YAML is shown in a **preview**; click **✎ Edit** to customise it (branches, Node version, extra steps…). Your edits apply to Copy, Download, and Write to repo. Use **↺ Reset** to regenerate.
317
+ - **Copy** / **Download** are always available.
318
+ - **Write to repo** (opt-in) writes the file into your project at its standard path (`.github/workflows/assuremind.yml`, `.gitlab-ci.yml`, or `Jenkinsfile`). Enable it in **Settings → CI/CD Integration → "Write CI files to repo"**. It only **writes the file** — it does **not** commit or push. Review the change and commit it from the **Git Control Center** when ready.
319
+
320
+ Exit code `0` = all passed · `1` = failures.
321
+
322
+ ---
323
+
324
+ ## Enterprise Features
325
+
326
+ ### Faker Data
327
+
328
+ 100+ @faker-js/faker generators across 16 categories generate realistic random data without code.
329
+
330
+ | Generator | Example output |
331
+ |-----------|---------------|
332
+ | `person.fullName` | Sarah Johnson |
333
+ | `internet.email` | sarah.johnson@gmail.com |
334
+ | `phone.number` | +1-555-123-4567 |
335
+ | `finance.amount` | 249.99 |
336
+ | `location.city` | San Francisco |
337
+ | Sequence `USER-{n}` | USER-001, USER-002, USER-003 |
338
+
339
+ Browse by category or search, select multiple tokens at once — selected tokens appear as removable chips, and clicking "Insert N tokens" inserts them all comma-separated (e.g. `{{FAKE_PERSON_FIRSTNAME}}, {{FAKE_PERSON_LASTNAME}}, {{FAKE_INTERNET_EMAIL}}`). Great for form filling. Tokens resolve to fresh random values on every test run — no more "email already exists" failures.
340
+
341
+ ### File Upload & Download
342
+
343
+ **Upload** — In the Test Editor, click the **Files** button (next to *Fake Data*) to upload a fixture file. It's saved in the repo at `fixtures/uploads/` (so it's committed and works in CI) and inserted into your step as a **`{{FILE:name}}`** token. The token resolves to the file's path at runtime:
344
+
345
+ ```
346
+ Step: Upload {{FILE:resume.pdf}} to the Resume field
347
+ Code: await page.getByLabel('Resume').setInputFiles('{{FILE:resume.pdf}}');
348
+ ```
349
+
350
+ For a file-chooser button, register the listener before the click (in one step):
351
+ ```js
352
+ const fileChooserPromise = page.waitForEvent('filechooser');
353
+ await page.getByRole('button', { name: 'Choose File' }).click();
354
+ (await fileChooserPromise).setFiles('{{FILE:resume.pdf}}');
355
+ ```
356
+
357
+ **Download** — Downloads are **captured automatically** during a run (no `saveAs`, no `fs` needed) and saved to `results/downloads/<runId>/`. The step waits until the file finishes downloading. After the run, open **Reports → Run Reports**, expand the run, and the **Downloads** panel lists every captured file — click any to save it to your machine.
358
+
359
+ **Example steps** (type these in plain English):
360
+
361
+ ```text
362
+ # Upload
363
+ Upload {{FILE:resume.pdf}} to the Resume field
364
+ Upload {{FILE:resume.pdf}} to the Resume field and wait for 'Upload complete' # most reliable
365
+ Upload {{FILE:photo.png}} # page's file input
366
+
367
+ # Download (file finishes downloading before the step passes)
368
+ Download the "Export CSV" link
369
+ Click the "Export" button to download
370
+ Download the file by clicking "Report PDF"
371
+ ```
372
+
373
+ Tip: for uploads, the `…and wait for '<success text>'` form is the most reliable (uploads have no completion event); otherwise the step waits for the network to settle.
374
+
375
+ ### Visual Regression
376
+
377
+ Pixel-perfect screenshot comparison catch UI changes that functional tests miss. Configure in **Settings** → Visual Regression. Review diffs in **Reports** → Visual Diffs tab.
378
+
379
+ 1. Enable in **Settings**Visual Regression, then toggle the **Eye icon** on any step in the Test Editor
380
+ 2. First run captures **baseline** screenshots (committed to git)
381
+ 3. Subsequent runs **compare** pixel-by-pixel using `pixelmatch`
382
+ 4. Diff > threshold → step **fails** with highlighted diff image
383
+ 5. Review in **Reports** → Visual Diffs → **Approve** (update baseline) or **Reject** (it's a bug)
384
+
385
+ ```
386
+ baselines/ # committed to git
387
+ ├── login-suite/login-test/
388
+ │ ├── step-3-chromium-1920x1080.png
389
+ │ └── step-3-firefox-1920x1080.png
390
+ results/visual-diffs/{runId}/ # gitignored, per-run
391
+ ├── step-3-baseline.png
392
+ ├── step-3-actual.png
393
+ └── step-3-diff.png # red highlights
394
+ ```
395
+
396
+ ### CI/CD Integration
397
+
398
+ Quality gates, PR comments, and notifications — configured from Studio, executed in pipelines.
399
+
400
+ | Feature | What it does |
401
+ |---------|-------------|
402
+ | **Quality gates** | Set min pass rate (95%), max duration, required tags — pipeline fails if gate fails |
403
+ | **PR comments** | Auto-post test results as GitHub PR comment or GitLab MR note |
404
+ | **Notifications** | Slack, Teams, Email, custom webhook — trigger on failure, every run, or healing events |
405
+
406
+ ```yaml
407
+ # GitHub Actions — full integration
408
+ - name: Run tests
409
+ env:
410
+ AI_PROVIDER: anthropic
411
+ ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
412
+ run: npx assuremind run --all --ci
413
+ # Quality gate auto-evaluates, PR comment auto-posts, Slack notifies on failure
414
+ ```
415
+
416
+ ---
417
+
418
+ ## Documentation
419
+
420
+ | Resource | Location |
421
+ |----------|----------|
422
+ | Getting started | [docs/GETTING-STARTED.md](docs/GETTING-STARTED.md) |
423
+ | Writing test steps (UI & API examples) | [docs/WRITING-STEPS.md](docs/WRITING-STEPS.md) |
424
+ | Studio walkthrough | [docs/STUDIO.md](docs/STUDIO.md) |
425
+ | CLI reference | [docs/CLI-REFERENCE.md](docs/CLI-REFERENCE.md) |
426
+ | Contributing | [CONTRIBUTING.md](CONTRIBUTING.md) |
427
+ | All AI providers | `.env.example` |
428
+ | Built-in docs | Studio → **Docs** page |
429
+ | Enterprise features | Studio → **Faker Data**, **CI/CD** pages; Visual Regression in **Settings** + **Reports** |
430
+
431
+ ---
432
+
433
+ ## Community & Support
434
+
435
+ Have a question, found a bug, or want to shape what comes next? Join the **[AssureMind Community Hub](https://assuremind.in/community/)**:
436
+
437
+ - 🐞 **Report bugs**
438
+ - ✨ **Request new features**
439
+ - 💡 **Share ideas and feedback**
440
+ - ❓ **Ask questions**
441
+
442
+ 🌐 **Website:** [https://assuremind.in/](https://assuremind.in/) &nbsp;|&nbsp; 💬 **Community Hub:** [https://assuremind.in/community/](https://assuremind.in/community/) &nbsp;|&nbsp; 📧 **[assuremind.contact@gmail.com](mailto:assuremind.contact@gmail.com)**
443
+
444
+ ---
445
+
446
+ ## License
447
+
448
+ Apache 2.0 — see [LICENSE](LICENSE) for details.
449
+
450
+ ---
451
+
452
+ *Built by [Deepak Hiremath](https://www.linkedin.com/in/deepak-hiremath-0017937a/)*