Deploy and Test LangGraph Workflow

Deploy your LangGraph application as a runnable service (e.g., using FastAPI or a similar framework). Create a comprehensive suite of test cases with varying levels of bias, factual inaccuracies, and compliance violations, then run them through your deployed system. Implement a logging mechanism to capture agent decisions and model outputs. Monitor the performance of your GPT-5 Pro and Gemini 3 Flash models using the monitoring tools provided by AI21 Studio and Together AI to identify any inference bottlenecks, latency issues, or unexpected model behaviors.
```python
from fastapi import FastAPI
from langchain_core.runnables import RunnableConfig
import uvicorn # Assuming 'app' is your compiled LangGraph application from Prompt 1
# app = graph_builder.compile() # fastapi_app = FastAPI(
# title="LangGraph Content Review API",
# description="API for autonomous content bias and fact-checking.",
# version="1.0.0",
# ) # @fastapi_app.post("/review_content/")
# async def review_content_endpoint(request: dict):
# content_text = request.get("text")
# if not content_text:
# return {"error": "'text' field is required"} # initial_state = {
# "messages": [("user", content_text)], # Initial user message can kick off workflow
# "content_to_review": content_text,
# "bias_analysis": "",
# "fact_check_results": [],
# "compliance_report": "",
# "has_issues": False
# } # # Invoke the LangGraph application
# config = RunnableConfig(recursion_limit=50) # Set recursion limit for graph traversal
# final_state = app.invoke(initial_state, config=config) # return {
# "status": "completed",
# "results": {
# "bias_analysis": final_state.get("bias_analysis"),
# "fact_check_results": final_state.get("fact_check_results"),
# "compliance_report": final_state.get("compliance_report"),
# "overall_issues_detected": final_state.get("has_issues")
# }
# } # To run the FastAPI app:
# if __name__ == "__main__":
# uvicorn.run(fastapi_app, host="0.0.0.0", port=8000) # Testing with curl (after running uvicorn):
# curl -X POST "http://localhost:8000/review_content/" -H "Content-Type: application/json" -d '{"text": "Your test article content here."}'
```