README.md

supadata-langchain

Supadata document loader integration for LangChain (Python).

This package provides a SupadataLoader that calls Supadata’s video endpoints:

• transcript — fetch a transcript for a social media video/post URL
• metadata — fetch structured metadata for a social media video/post URL

Supadata supports social media video/post URLs (YouTube, TikTok, Instagram, Facebook, and X/Twitter).

Usage

import os
from supadata_langchain import SupadataLoader

os.environ["SUPADATA_API_KEY"] = "YOUR_API_KEY"

loader = SupadataLoader()

docs = loader.load(
    {
        "url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
        "operation": "transcript",
        "lang": "en",
        "text": True,
        "mode": "auto",
    }
)

print(docs[0].page_content)
print(docs[0].metadata)

Metadata

docs = loader.load(
    {
        "url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
        "operation": "metadata",
    }
)

print(docs[0].page_content)

API

SupadataLoader(api_key: str | None = None, base_url: str = "https://api.supadata.ai")

• api_key defaults to the SUPADATA_API_KEY environment variable.

load(params: dict) -> list[Document]

params supports:

• url (string, required)
• operation ("transcript" or "metadata", default "transcript")
• lang (string, optional; transcript only)
• text (bool, optional; transcript only)
• mode ("auto" | "native" | "generate", optional; transcript only)

Returns a list containing a single Document.