Fix FastAPI async OpenAI error
Quick answer
FastAPI async errors with
OpenAI occur because the official OpenAI Python SDK is synchronous and not awaitable. Use synchronous calls inside FastAPI endpoints or run the SDK calls in a thread executor to fix the error. ERROR TYPE
code_error ⚡ QUICK FIX
Call
OpenAI client methods synchronously inside FastAPI or use asyncio.to_thread to run blocking calls without blocking the event loop.Why this happens
The official OpenAI Python SDK methods like client.chat.completions.create() are synchronous blocking calls. When you call them directly inside an async def FastAPI route handler using await, Python raises a TypeError because the method is not awaitable.
Example error output:
TypeError: object ChatCompletionCreate is not awaitable
Typical broken code:
from fastapi import FastAPI
from openai import OpenAI
import os
app = FastAPI()
client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])
@app.get("/chat")
async def chat():
# Incorrect: calling synchronous method with await
response = await client.chat.completions.create(
model="gpt-4o",
messages=[{"role": "user", "content": "Hello"}]
)
return {"reply": response.choices[0].message.content} output
TypeError: object ChatCompletionCreate is not awaitable
The fix
Use the OpenAI client synchronously inside FastAPI route handlers without await. Since FastAPI supports sync route handlers, you can make the route a normal def function or run the blocking call in a thread executor to keep the route async.
This example uses asyncio.to_thread to run the blocking call without blocking the event loop:
import asyncio
from fastapi import FastAPI
from openai import OpenAI
import os
app = FastAPI()
client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])
@app.get("/chat")
async def chat():
response = await asyncio.to_thread(
client.chat.completions.create,
model="gpt-4o",
messages=[{"role": "user", "content": "Hello"}]
)
return {"reply": response.choices[0].message.content} output
{"reply": "Hello! How can I assist you today?"} Preventing it in production
To avoid async errors in production FastAPI apps using the synchronous OpenAI SDK:
- Use
asyncio.to_threadorrun_in_executorto run blocking calls without blocking the event loop. - Alternatively, define your FastAPI route handlers as synchronous
deffunctions if you do not need async concurrency. - Implement retry logic with exponential backoff for transient API errors.
- Validate API keys and handle exceptions gracefully to prevent crashes.
Key Takeaways
- The OpenAI Python SDK is synchronous; do not await its methods directly in async FastAPI routes.
- Use asyncio.to_thread to run blocking OpenAI calls without blocking FastAPI's event loop.
- Define FastAPI routes as synchronous if you do not require async concurrency for OpenAI calls.