Add YouTube transcript tool (although not very helpful)

agent.py

@@ -1,7 +1,7 @@
 import logfire
 
 from pydantic_ai import Agent
-from tools.safe_duck import safe_duckduckgo_search_tool
+from tools import safe_duckduckgo_search_tool, get_youtube_transcript
 logfire.configure()  
 logfire.instrument_pydantic_ai()
 
@@ -9,7 +9,7 @@ class BasicAgent:
     def __init__(self):
         self.agent = Agent(
             "openai:o3-mini",
-            tools=[safe_duckduckgo_search_tool()],
+            tools=[safe_duckduckgo_search_tool(), get_youtube_transcript],
             system_prompt="Search DuckDuckGo for the given query and return the results.",
         )
 

requirements.lock

@@ -50,6 +50,8 @@ colorama==0.4.6
     # via griffe
 cryptography==45.0.2
     # via authlib
+defusedxml==0.7.1
+    # via youtube-transcript-api
 deprecated==1.2.18
     # via
     #   opentelemetry-api
@@ -299,6 +301,7 @@ requests==2.32.3
     #   huggingface-hub
     #   opentelemetry-exporter-otlp-proto-http
     #   pydantic-ai-slim
+    #   youtube-transcript-api
 rich==14.0.0
     # via
     #   logfire
@@ -397,5 +400,7 @@ wrapt==1.17.2
     # via
     #   deprecated
     #   opentelemetry-instrumentation
+youtube-transcript-api==1.0.3
+    # via -r requirements.txt
 zipp==3.21.0
     # via importlib-metadata

requirements.txt

@@ -3,4 +3,5 @@ gradio[oauth]
 requests
 pydantic-ai
 pydantic-ai-slim[duckduckgo]
-pydantic-ai[logfire]
+pydantic-ai[logfire]
+youtube-transcript-api

tools/__init__.py

@@ -0,0 +1,2 @@
+from .safe_duck import safe_duckduckgo_search_tool
+from .youtube_transcript_tool import get_youtube_transcript 

tools/youtube_transcript_tool.py

@@ -0,0 +1,19 @@
+from youtube_transcript_api import YouTubeTranscriptApi
+
+def get_youtube_transcript(video_id: str) -> str:
+    """
+    Fetches the transcript for a given YouTube video ID.
+
+    Args:
+        video_id: The ID of the YouTube video.
+
+    Returns:
+        The transcript of the video as a string, or an error message if the transcript cannot be fetched.
+    """
+    try:
+        transcript_list = YouTubeTranscriptApi.list_transcripts(video_id)
+        transcript = transcript_list.find_generated_transcript(['en'])
+        fetched_transcript = transcript.fetch()
+        return " ".join([segment['text'] for segment in fetched_transcript])
+    except Exception as e:
+        return f"Error fetching transcript: {str(e)}"