CreatorOS Python SDK

Official Python SDK for CreatorOS — High-throughput Web-to-Markdown extraction, autonomous multi-URL agent search, and studio-grade AI text-to-speech.

Zero external dependencies. Works on standard Python 3.8+ (urllib based, zero conflicts with Pydantic/Torch/LangChain).

Installation

pip install creatoros

Quickstart

1. Instant Web-to-Markdown Extraction

Extract noise-free Markdown with article metadata stripped of navigation, ads, and tracker boilerplate:

import creatoros

# Free tier (no key required for up to 20 daily calls)
markdown = creatoros.extract("https://news.ycombinator.com")
print(markdown)

2. Autonomous Agent Live Web Search

Search the live web and retrieve structured results with pre-extracted content from top matches:

import creatoros

results = creatoros.search("FastAPI tutorial 2026", limit=3)
for hit in results.get("results", []):
    print(f"[{hit['rank']}] {hit['title']} -> {hit['url']}")

3. Neural English Text-to-Speech

Generate human-cadence English speech:

import creatoros

audio_bytes = creatoros.speak(
    text="Welcome to CreatorOS. Clean markdown extraction for autonomous agents.",
    voice="anna",
    output_file="welcome.wav"
)

Using with Developer Passes (High Concurrency)

If you have an active CreatorOS Developer Pass, provide your key via parameter or the CREATOROS_API_KEY environment variable:

import os
from creatoros import CreatorOS

client = CreatorOS(api_key="YOUR_CREATOROS_PASS_KEY")

# High-concurrency extraction (up to 500,000 monthly credits)
doc = client.extract("https://example.com")

LangChain Integration

CreatorOS provides drop-in tools compatible with LangChain, AutoGen, and CrewAI:

from creatoros.langchain import CreatorOSExtractTool, CreatorOSSearchTool

extract_tool = CreatorOSExtractTool()
search_tool = CreatorOSSearchTool()

# Use inside LangChain Agent
tools = [extract_tool, search_tool]

License

MIT © CreatorOS