fix: normalize search_volume (int→str), relevance (float→int), competition (lower→Title), post_count (int→str) for LongCat API responses

fix: add catch-all viral_hashtags mapping at end of cross_map_fields to catch hashtags set by platform-specific mappers

fix: add generic cross-map fields (core_keywords, viral_hashtags, strategy_tips, content_titles) to all platform response models so they survive FastAPI's response_model filter

fix: add Pinterest cross-maps for core_keywords, content_titles, hashtags, strategy_tips

fix: handle model returning dict instead of list for tweet_threads/post_drafts/post_ideas/video_concepts

feat: Add LinkedIn API posting + enhanced analyzer + schema

feat: Add LongCat API as primary model with Qwen fallback + cross-map fix

fix: Add Instagram cross-maps for hashtag_sets/captions/content_ideas

fix: Run cross-map after defaults so LLM-sparse responses still get mapped

fix: Add core_keywords and related_phrases cross-mapping from tags/video_titles

fix: Cross-map platform keys to generic keys for frontend template

fix: Improve prompts to return JSON object, add error for list response

fix: Add error message when LLM returns list instead of dict

fix: Escape curly braces in f-string causing Invalid format specifier

fix: Guard against LLM returning JSON array instead of object

fix: Add missing validate_and_fill_data_defaults function

fix: Catch normalize/validate errors, hide raw JSON parse errors againe

fix: Catch normalize/validate errors, hide raw JSON parse errors

feat: Multi-platform batch generation + simple UI

feat: Simple HTML frontend with section copy buttons

fix: Handle missing correction fields + ResponseValidationError

feat: Add Grammar Correction API v2.1.0

fix: Add normalize_response to handle model output format mismatches

perf: Reduce max_new_tokens 1500→800, cut output by 50% for faster CPU inference

fix: Raise HTTPException for unexpected errors instead of returning invalid dict

feat: Add 3-tier quantization with auto hardware detection

fix: Simplify prompts for 0.5B model reliability + resilient error handling

feat: Add multi-platform social media analysis with 7 platform-specific endpoints

Add GPU Setup Cheat Sheet for RTX 4050

Add documentation for running on RTX 4050 6GB

Update docs to reflect Qwen-0.5B Local CPU architecture

Add Schema Validation Safety Net: Auto-fill missing JSON fields with defaults

Fix JSON Crash: Lower Temp to 0.3 + Enhance Repair Logic + Add Error Fallback

Switch to Qwen-0.5B (Nano) for instant speed and 100% stability

Optimize for Speed: 30 Keywords / 1500 Tokens to stop Local CPU Timeout

Reduce to 50 keywords to prevent CPU Timeout (Internal Server Error)

Scale up to 150 keywords using Local Inference

Migrate to Local Inference (Qwen-1.5B) to permanently bypass API limits: Re-added Torch/Transformers

Switch to SmolLM2 (HF Native) to resolve provider support issues

Switch to Qwen-1.5B (Lightweight) to fix all API limits

Switch to Google Gemma-2-9B-IT for dependable serverless support

Fix NameError and Switch to Phi-3-Mini to bypass 402 Payment Limits

Fix truncation error: Reduce to 50 keywords + JSON repair logic

Increase keyword limit to 150 and max_tokens to 8000

Extend output: 10 keywords/hashtags + add Content Titles

Update docs with A-Z guide for Qwen Model

Update prompts for stronger SEO data extraction

Switch to Qwen/Qwen2.5-7B-Instruct for reliable availability

Switch to Zephyr-7B-Beta for Chat API compatibility

Fix API method: Use chat_completion for Mistral Instruct