w

main.py

@@ -33,29 +33,34 @@ async def get_transcript(req: TranscriptRequest):
     if not video_id:
         raise HTTPException(status_code=400, detail="Invalid YouTube URL")
 
-    # We will try multiple Invidious instances in case one is down
+    # Using Piped API - much more reliable on cloud servers than Invidious
     instances = [
-        "https://inv.nadeko.net",
-        "https://invidious.fdn.fr",
-        "https://invidious.protokolla.fi",
-        "https://iv.datura.network"
+        "https://pipedapi.kavin.rocks",
+        "https://pipedapi.adminforge.de",
+        "https://piped-api.privacy.com.de",
+        "https://api.piped.projectsegfau.lt"
     ]
 
     for host in instances:
         try:
-            print(f"Trying {host} for video {video_id}...")
-            # 1. Fetch available captions from Invidious
-            res = requests.get(f"{host}/api/v1/captions/{video_id}?label=English", timeout=5)
+            print(f"Trying Piped instance {host} for video {video_id}...")
+            
+            # 1. Fetch video stream data (contains caption list)
+            res = requests.get(f"{host}/streams/{video_id}", timeout=8)
             
             if res.status_code != 200:
-                continue # Try next instance
+                print(f"Failed on {host}: Status {res.status_code}")
+                continue
                 
             data = res.json()
-            if not data.get("captions"):
-                continue # Try next instance
+            captions = data.get("captions", [])
             
-            # 2. Find the English caption URL (or first available)
-            caption = next((c for c in data["captions"] if "English" in c.get("label", "")), data["captions"][0])
+            if not captions:
+                print(f"No captions on {host}")
+                continue
+            
+            # 2. Find English caption (or first available)
+            caption = next((c for c in captions if c.get("code") == "en" or "English" in c.get("name", "")), captions[0])
             caption_url = caption.get("url")
             
             if not caption_url:
@@ -64,27 +69,37 @@ async def get_transcript(req: TranscriptRequest):
             if caption_url.startswith("/"):
                 caption_url = f"{host}{caption_url}"
 
-            # 3. Fetch the actual caption XML
-            cap_res = requests.get(caption_url, timeout=5)
+            print(f"Fetching caption data from {caption_url}...")
+            
+            # 3. Fetch the JSON caption data
+            cap_res = requests.get(caption_url, timeout=8)
             if cap_res.status_code != 200:
                 continue
             
-            # 4. Parse XML to plain text
-            texts = re.findall(r'<text[^>]*>([^<]*)</text>', cap_res.text)
+            cap_data = cap_res.json()
+            
+            # 4. Parse Piped JSON captions into plain text
+            texts = []
+            for event in cap_data.get("events", []):
+                for seg in event.get("segs", []):
+                    text = seg.get("utf8", "").strip()
+                    if text and text != "\n":
+                        texts.append(text)
+            
             if not texts:
                 continue
             
             full_text = " ".join(texts)
-            full_text = full_text.replace("&amp;#39;", "'").replace("&amp;quot;", '"').replace("&amp;amp;", "&").replace("&#39;", "'").replace("\n", " ")
+            full_text = full_text.replace("&amp;#39;", "'").replace("&amp;quot;", '"').replace("&amp;amp;", "&").replace("&#39;", "'")
             
             print(f"Successfully fetched transcript from {host}. Length: {len(full_text)} chars")
             return {"text": full_text}
             
         except Exception as e:
             print(f"Failed on {host}: {e}")
-            continue # Try next instance
-            
-    raise HTTPException(status_code=500, detail="Failed to fetch transcript. All sources timed out or are blocked.")
+            continue
+
+    raise HTTPException(status_code=500, detail="Failed to fetch transcript. All sources timed out or video has no captions.")
 
 # ========================================================
 # TRANSLATE API