FastAPI expects the body of a POST request to follow a structure.
If you send just a plain string like:

"hello world"

FastAPI cannot parse it unless you tell it exactly how to treat that string.

So it returns something like:

{"detail":"There was an error parsing the body"}

or:

{"detail":[{"type":"string_type","msg":"str type expected"}]}

✅ How to Fix It

✔ Fix 1: Use a Pydantic Model (Recommended)

from fastapi import FastAPI
from pydantic import BaseModel

app = FastAPI()

class TextIn(BaseModel):
    text: str

@app.post("/predict")
def predict(data: TextIn):
    input_text = data.text
    # your ML model prediction here
    result = my_model.predict([input_text])
    return {"prediction": result}

Then send this JSON body:

{
  "text": "This is my input sentence"
}