Upload ai_ext.py

Upload ai_ext.py

Upload ai_ext.py

Upload ai_ext.py

Upload ai_ext.py with huggingface_hub

Delete *

Restore to commit 6c739c9

Restore to commit 6c739c9

probe: more candidate models + full error

Add /api/ai/probe diagnostic to find fast available chat model

Cache last-working LLM to avoid retrying unavailable models (faster subsequent rewrites)

Speed up rewrite: use fast text LLM for summaries (no vision needed), text-only for rewrite/url/topic, 60s timeout

Speed up rewrite: background short-video generation; topic_post text-only Qwen (no image input)

Add emotion (cảm xúc) TTS + AI auto-select voice+emotion per topic

Fix: multi-segment short video - each key point shown with its own image

Filter TTS voices: VN (HoaiMy/NamMinh) + gTTS + 12 Multilingual only

Fix ai_ext.py - restore from c93b544

Test upload size - first half of ai_ext.py from c93b544

Restore: ai_ext.py, ai_fix2.py, ai_patch.py from c93b544

Delete restore_runner.py, patch_runtime.py, patch_extra.py, app_entry.py, app_clean.py, wc2026_scraper.py, ai_ext.py, ai_fix2.py, ai_patch.py, app_v2_entry.py, Dockerfile, .gitattributes

Restore ai_ext.py from c93b544 - AI extension endpoints

Upload ai_ext.py

Fix rewrite: dedup sentences, complete sentence trim, merge images from direct+Jina scrape (#1)

Update ai_ext.py: stricter rewrite, dedup, all images, 24 multilingual voices

Delete ai_ext.py

Upload ai_ext.py

Upload ai_ext.py

Upload ai_ext.py

Restore Space to revision 33c3dda

Restore ai_ext.py from 33c3dda

Summaries only and add AI shorts video generation from wall posts

Change AI rewrite to concise main-point summaries only

Remove AI compose box and fix HF token/Qwen/Jina handling

Fix AI scraping with Jina Reader/Search and require Qwen for summaries

Fix AI URL input persistence and improve internet-sourced topic posts

Add AI wall topic/url Qwen2.5-VL extension