feat: Brainstorm Topics with GSC + Issue #518 fixes + Blog Editor enhancements

Issue #518 - Subscription not updating after checkout:
- Fix stale closure in SubscriptionContext checkout polling (use subscriptionRef)
- Move checkout success polling from InitialRouteHandler into SubscriptionContext
- Remove redundant polling code from InitialRouteHandler
- Fix plan label: 'Free' instead of 'No Plan', proper capitalization
- Add plan refresh button in UserBadge
- Add 'View Costing Details' to UserBadge dropdown
- Rename 'ALwrity Podcast Maker' to 'Podcast Creator' across UI
- Clean subscription=success URL param after verification

Blog Writer WYSIWYG Editor enhancements:
- Per-section preview toggle (view/edit icons)
- Enhanced hover-based toolbar
- Circular SVG progress stats bar with detailed tooltip
- Research tool chips in stats bar footer
- Per-section TTS with useTextToSpeech hook (browser native)
- Full blog preview modal with print/PDF support
- PlayAllTTSButton: sequential playback with progress bar
- OnThisPageNav: floating sidebar with scroll tracking
- Section data attributes for scroll anchoring

GSC Brainstorm Topics feature:
- Backend: gsc_brainstorm_service.py (rule-based + LLM recommendations)
- Backend: POST /gsc/brainstorm endpoint with 3-word minimum validation
- Frontend: gscBrainstorm.ts API client
- Frontend: useGSCBrainstormConnection hook (popup OAuth, no /onboarding redirect)
- Frontend: useGSCBrainstorm hook (connect check + brainstorm call)
- Frontend: GSCBrainstormModal (3-tab results: Opportunities, Gaps, AI Recs)
- Frontend: BrainstormButton (visible at 3+ words, GSC connect overlay)
- Wire BrainstormButton into ManualResearchForm and ResearchAction
- Add blog_writer to gsc_auth router features for ALWRITY_ENABLED_FEATURES

backend/services/llm_providers/main_text_generation.py

@@ -429,6 +429,23 @@ def llm_text_gen(
         except Exception as provider_error:
             logger.error(f"[llm_text_gen] Provider {gpt_provider} failed: {str(provider_error)}")
             
+            # Surface balance/quota errors immediately without fallback
+            error_str = str(provider_error).lower()
+            if "insufficient_balance" in error_str or "balance_not_enough" in error_str or ("403" in error_str and "balance" in error_str):
+                logger.error(f"[llm_text_gen] Balance/quota error from {gpt_provider}, not attempting fallback")
+                raise HTTPException(
+                    status_code=403,
+                    detail={
+                        "error": "insufficient_balance",
+                        "message": f"Your {gpt_provider.capitalize()} API balance is insufficient. Please top up your account or switch providers.",
+                        "usage_info": {
+                            "error_type": "insufficient_balance",
+                            "provider": gpt_provider,
+                            "suggestion": f"Set GPT_PROVIDER=google in your environment to use Gemini instead, or add credits to your {gpt_provider.capitalize()} account."
+                        }
+                    }
+                )
+            
             # CIRCUIT BREAKER: Only try ONE fallback to prevent expensive API calls
             fallback_providers = ["google", "huggingface"]
             fallback_providers = [p for p in fallback_providers if p in available_providers and p != gpt_provider]