academia-mcp 1.10.6__tar.gz → 1.10.8__tar.gz

{academia_mcp-1.10.6 → academia_mcp-1.10.8}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: academia-mcp
-Version: 1.10.6
+Version: 1.10.8
 Summary: MCP server that provides different tools to search for scientific publications
 Author-email: Ilya Gusev <phoenixilya@gmail.com>
 Project-URL: Homepage, https://github.com/IlyaGusev/academia_mcp
@@ -30,6 +30,7 @@ Requires-Dist: datasets>=4.0.0
 Requires-Dist: pymupdf>=1.26.4
 Requires-Dist: pillow>=11.3.0
 Requires-Dist: pydantic-settings>=2.6.0
+Requires-Dist: youtube-transcript-api>=1.2.2
 Dynamic: license-file
 
 # Academia MCP

{academia_mcp-1.10.6 → academia_mcp-1.10.8}/academia_mcp/server.py

@@ -42,6 +42,7 @@ from academia_mcp.tools.bitflip import (
 from academia_mcp.tools.review import review_pdf_paper, download_pdf_paper
 from academia_mcp.tools.show_image import show_image, describe_image
 from academia_mcp.tools.speech_to_text import speech_to_text
+from academia_mcp.tools.yt_transcript import yt_transcript
 
 
 def configure_uvicorn_style_logging(level: int = logging.INFO) -> None:
@@ -92,6 +93,7 @@ def run(
     server.add_tool(get_latex_templates_list)
     server.add_tool(visit_webpage)
     server.add_tool(show_image)
+    server.add_tool(yt_transcript)
 
     if settings.WORKSPACE_DIR:
         server.add_tool(compile_latex)

{academia_mcp-1.10.6 → academia_mcp-1.10.8}/academia_mcp/settings.py

@@ -21,6 +21,9 @@ class Settings(BaseSettings):
     DOCUMENT_QA_DOCUMENT_MAX_LENGTH: int = 200000
     DESCRIBE_IMAGE_MODEL_NAME: str = "gpt-4.1"
 
+    WEBSHARE_PROXY_USERNAME: Optional[str] = None
+    WEBSHARE_PROXY_PASSWORD: Optional[str] = None
+
     PORT: int = 5056
     WORKSPACE_DIR: Optional[Path] = None
 

{academia_mcp-1.10.6 → academia_mcp-1.10.8}/academia_mcp/tools/__init__.py

@@ -16,6 +16,7 @@ from .bitflip import extract_bitflip_info, generate_research_proposals, score_re
 from .review import review_pdf_paper, download_pdf_paper, review_pdf_paper_by_url
 from .show_image import show_image, describe_image
 from .speech_to_text import speech_to_text
+from .yt_transcript import yt_transcript
 
 __all__ = [
     "arxiv_search",
@@ -45,4 +46,5 @@ __all__ = [
     "show_image",
     "describe_image",
     "speech_to_text",
+    "yt_transcript",
 ]

{academia_mcp-1.10.6 → academia_mcp-1.10.8}/academia_mcp/tools/web_search.py

@@ -1,5 +1,5 @@
 import json
-from typing import Optional, List
+from typing import Optional, List, Tuple
 
 from academia_mcp.utils import post_with_retries, get_with_retries
 from academia_mcp.settings import settings
@@ -12,6 +12,24 @@ BRAVE_SEARCH_URL = "https://api.search.brave.com/res/v1/web/search"
 EXCLUDE_DOMAINS = ["chatpaper.com"]
 
 
+def _parse_domains(query: str) -> Tuple[str, List[str]]:
+    site_term = "site:"
+    if site_term not in query:
+        return query, []
+    parts = query.split()
+    query_parts = []
+    include_domains = []
+    for part in parts:
+        if not part.startswith(site_term):
+            query_parts.append(part)
+            continue
+        domain = part[len(site_term) :]
+        if domain:
+            include_domains.append(domain)
+    query = " ".join(query_parts)
+    return query, include_domains
+
+
 def web_search(
     query: str,
     limit: Optional[int] = 20,
@@ -40,6 +58,13 @@ def web_search(
             isinstance(domain, str) for domain in include_domains
         ), "Error: include_domains should be a list of strings"
 
+    query, query_include_domains = _parse_domains(query)
+    if query_include_domains:
+        if include_domains:
+            include_domains.extend(query_include_domains)
+        else:
+            include_domains = query_include_domains
+
     is_tavily_available = bool(settings.TAVILY_API_KEY)
     is_exa_available = bool(settings.EXA_API_KEY)
     is_brave_available = bool(settings.BRAVE_API_KEY)

academia_mcp-1.10.8/academia_mcp/tools/yt_transcript.py

@@ -0,0 +1,30 @@
+from youtube_transcript_api import YouTubeTranscriptApi
+from youtube_transcript_api.proxies import WebshareProxyConfig
+
+from academia_mcp.settings import settings
+
+
+def yt_transcript(video_url: str) -> str:
+    """
+    Tool to fetch the transcript of a YouTube video given its URL.
+
+    Returns a transcript of the video as a single string.
+
+    Args:
+        video_url (str): YouTube video URL.
+    """
+    if "youtu.be" in video_url:
+        video_id = video_url.strip().split("youtu.be/")[1]
+    else:
+        video_id = video_url.strip().split("v=")[-1]
+    video_id = video_id.split("?")[0]
+    proxy_config = None
+    if settings.WEBSHARE_PROXY_USERNAME and settings.WEBSHARE_PROXY_PASSWORD:
+        proxy_config = WebshareProxyConfig(
+            proxy_username=settings.WEBSHARE_PROXY_USERNAME,
+            proxy_password=settings.WEBSHARE_PROXY_PASSWORD,
+        )
+    api = YouTubeTranscriptApi(proxy_config=proxy_config)
+    transcript = api.fetch(video_id)
+    snippets = transcript.snippets
+    return "\n".join([f"{int(entry.start)}: {' '.join(entry.text.split())}" for entry in snippets])

{academia_mcp-1.10.6 → academia_mcp-1.10.8}/academia_mcp.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: academia-mcp
-Version: 1.10.6
+Version: 1.10.8
 Summary: MCP server that provides different tools to search for scientific publications
 Author-email: Ilya Gusev <phoenixilya@gmail.com>
 Project-URL: Homepage, https://github.com/IlyaGusev/academia_mcp
@@ -30,6 +30,7 @@ Requires-Dist: datasets>=4.0.0
 Requires-Dist: pymupdf>=1.26.4
 Requires-Dist: pillow>=11.3.0
 Requires-Dist: pydantic-settings>=2.6.0
+Requires-Dist: youtube-transcript-api>=1.2.2
 Dynamic: license-file
 
 # Academia MCP

{academia_mcp-1.10.6 → academia_mcp-1.10.8}/academia_mcp.egg-info/SOURCES.txt

@@ -33,6 +33,7 @@ academia_mcp/tools/show_image.py
 academia_mcp/tools/speech_to_text.py
 academia_mcp/tools/visit_webpage.py
 academia_mcp/tools/web_search.py
+academia_mcp/tools/yt_transcript.py
 tests/test_anthology_search.py
 tests/test_arxiv_download.py
 tests/test_arxiv_search.py
@@ -46,4 +47,5 @@ tests/test_s2.py
 tests/test_show_image.py
 tests/test_speech_to_text.py
 tests/test_visit_webpage.py
-tests/test_web_search.py
+tests/test_web_search.py
+tests/test_yt_transcript.py

{academia_mcp-1.10.6 → academia_mcp-1.10.8}/academia_mcp.egg-info/requires.txt

@@ -18,3 +18,4 @@ datasets>=4.0.0
 pymupdf>=1.26.4
 pillow>=11.3.0
 pydantic-settings>=2.6.0
+youtube-transcript-api>=1.2.2

{academia_mcp-1.10.6 → academia_mcp-1.10.8}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "academia-mcp"
-version = "1.10.6"
+version = "1.10.8"
 description = "MCP server that provides different tools to search for scientific publications"
 readme = "README.md"
 authors = [
@@ -37,6 +37,7 @@ dependencies = [
     "pymupdf>=1.26.4",
     "pillow>=11.3.0",
     "pydantic-settings>=2.6.0",
+    "youtube-transcript-api>=1.2.2",
 ]
 
 [dependency-groups]

{academia_mcp-1.10.6 → academia_mcp-1.10.8}/tests/test_web_search.py

@@ -42,3 +42,14 @@ def test_web_search_include_domains() -> None:
     assert results
     assert len(results["results"]) > 0
     assert all("wikipedia.org" in result["url"] for result in results["results"])
+
+
+def test_web_search_include_query_domains() -> None:
+    results = web_search(
+        "site:wikipedia.org autoregressive models path-star graphs",
+    )
+    assert results
+    results = json.loads(results)
+    assert results
+    assert len(results["results"]) > 0
+    assert all("wikipedia.org" in result["url"] for result in results["results"])

academia_mcp-1.10.8/tests/test_yt_transcript.py

@@ -0,0 +1,13 @@
+from academia_mcp.tools import yt_transcript
+
+
+def test_yt_transcript_base() -> None:
+    result = yt_transcript("https://www.youtube.com/watch?v=21EYKqUsPfg")
+    assert result is not None
+    assert "chatting with richard sutton" in result.lower()
+
+
+def test_yt_transcript_short_link() -> None:
+    result = yt_transcript("https://youtu.be/21EYKqUsPfg?si=iity_X55GIWUQWuT")
+    assert result is not None
+    assert "chatting with richard sutton" in result.lower()