What is LlamaIndex?
Data framework for building RAG pipelines and LLM applications over custom data. Connects LLMs to external data sources.
Searching YouTube Playlists with LlamaIndex
This integration lets your LlamaIndex agent search YouTube Playlists in real time via the Scavio API. The agent gets back structured JSON with playlist entries from search, video results, channel results, shorts results -- ready for reasoning and decision-making.
Setup
Bash
pip install llama-index requestsCode Example
Here is a complete LlamaIndex agent that searches YouTube Playlists using Scavio:
Python
from llama_index.core.tools import FunctionTool
from llama_index.llms.openai import OpenAI
from llama_index.core.agent import ReActAgent
import json
import os
import requests
# Playlist data is not available: POST /api/v1/youtube/search accepts type: playlist but the provider returns an empty playlists array, so there is no playlist title, video count, creator or visibility to read, and no removal tracking. This sample returns the videos for the query; type: channel does work if you want creators instead.
def search_youtube_playlists(query: str) -> str:
"""Query YouTube Playlists through the Scavio API (POST /api/v1/youtube/search)."""
response = requests.post(
"https://api.scavio.dev/api/v1/youtube/search",
headers={
"Authorization": "Bearer " + os.environ["SCAVIO_API_KEY"],
"Content-Type": "application/json",
},
json={"search": query},
timeout=60,
)
response.raise_for_status()
return json.dumps(response.json())
tool = FunctionTool.from_defaults(fn=search_youtube_playlists)
llm = OpenAI(model="gpt-5.5")
agent = ReActAgent.from_tools([tool], llm=llm, verbose=True)
response = agent.chat("best rag tutorials playlist")
print(response)Full Working Example
A production-ready example with error handling:
Python
from llama_index.core.tools import FunctionTool
from llama_index.llms.openai import OpenAI
from llama_index.core.agent import ReActAgent
import json
import os
import requests
# Playlist data is not available: POST /api/v1/youtube/search accepts type: playlist but the provider returns an empty playlists array, so there is no playlist title, video count, creator or visibility to read, and no removal tracking. This sample returns the videos for the query; type: channel does work if you want creators instead.
def search_youtube_playlists(query: str) -> str:
"""Query YouTube Playlists through the Scavio API (POST /api/v1/youtube/search)."""
response = requests.post(
"https://api.scavio.dev/api/v1/youtube/search",
headers={
"Authorization": "Bearer " + os.environ["SCAVIO_API_KEY"],
"Content-Type": "application/json",
},
json={"search": query},
timeout=60,
)
response.raise_for_status()
return json.dumps(response.json())
tool = FunctionTool.from_defaults(fn=search_youtube_playlists)
llm = OpenAI(model="gpt-5.5")
agent = ReActAgent.from_tools([tool], llm=llm, verbose=True)
response = agent.chat("best rag tutorials playlist")
print(response)Pricing
Scavio offers a free tier with 50 credits on signup (1 credit per search). No credit card required. This is enough to build and test your LlamaIndex integration. Paid plans start at $30/month for higher volumes.