docs: align test count in README and document .env.example

- Update README test comment from "400+ tests" to "408 tests" to match badge
- Add grouped comments to .env.example explaining each variable's purpose and whether it is required or optional

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: lijunliu-gh

.env.example

@@ -1,8 +1,22 @@
+# ── LLM Provider (at least one required) ──────────────────────────
+# OpenAI or any OpenAI-compatible endpoint (Qwen, DeepSeek, GLM, MiniMax)
 OPENAI_API_KEY=
 OPENAI_MODEL=gpt-4.1-mini
 OPENAI_BASE_URL=
+
+# OpenRouter (optional, alternative multi-model gateway)
 OPENROUTER_API_KEY=
-TAVILY_API_KEY=
+
+# Anthropic Claude (optional, auto-detected by model name prefix "claude")
 ANTHROPIC_API_KEY=
+
+# Google Gemini (optional, auto-detected by model name prefix "gemini")
 GOOGLE_API_KEY=
+
+# ── Optional Services ─────────────────────────────────────────────
+# Tavily web search (optional, used for URL source fetching)
+TAVILY_API_KEY=
+
+# ── Output ────────────────────────────────────────────────────────
+# Directory for generated deck JSON and PPTX files (default: output)
 AUTO_PPT_OUTPUT_DIR=output

README.md

@@ -157,7 +157,7 @@ docker run --rm -it -e OPENAI_API_KEY auto-ppt-engine python mcp_server.py  # MC
 ## Testing
 
 ```bash
-python -m pytest tests/ -v   # 400+ tests
+python -m pytest tests/ -v   # 408 tests
 npm run ci:smoke             # JS renderer + end-to-end smoke
 ```