Discord机器人GPT斜杠命令报错：NotFound 404(10062) Unknown interaction

问题场景 #

我编写了一段Discord机器人的GPT斜杠命令代码，具体实现如下：

@bot.tree.command(description="create a txt with chat-gpt")
async def gpt(interaction: discord.Interaction, *, prompt:str):
    async with aiohttp.ClientSession() as session:
        payload = {
            "model":"text-davinci-003",
            "prompt":prompt,
            "temperature": 0.5,
            "max_tokens": 50,
            "presence_penalty": 0,
            "frequency_penalty": 0,
            "best_of": 1,
        }
        headers = {"Authorization": f"Bearer {API_KEY}"}
        async with session.post("https://api.openai.com/v1/completions", json=payload, headers=headers) as resp:
            response = await resp.json()
            em = discord.Embed(title="Chat gpt’s responce:", description=response)
            await interaction.response.defer()
            await asyncio.sleep(delay=0)
            await interaction.followup.send(embed=em)

执行该命令时触发如下错误：

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/home/haoroux/.local/lib/python3.10/site-packages/discord/app_commands/tree.py", line 1242, in _call
    await command._invoke_with_namespace(interaction, namespace)
  File "/home/haoroux/.local/lib/python3.10/site-packages/discord/app_commands/commands.py", line 887, in _invoke_with_namespace
    return await self._do_call(interaction, transformed_values)
  File "/home/haoroux/.local/lib/python3.10/site-packages/discord/app_commands/commands.py", line 880, in _do_call
    raise CommandInvokeError(self, e) from e
discord.app_commands.errors.CommandInvokeError: Command 'gpt' raised an exception: NotFound: 404 Not Found (error code: 10062): Unknown interaction

我尝试过调整延迟、将embed参数放入response.defer()、覆盖embed等方法，但问题仍未解决。

问题原因与修复方案 #

核心原因 #

Discord要求斜杠命令必须在3秒内给出初始响应（包括defer操作），否则会判定交互超时，后续的followup.send就会触发Unknown interaction错误。你的代码是先调用OpenAI API获取结果，再执行defer，这时候大概率已经超过了3秒的时间窗口。

修复步骤 #

1. 优先执行defer操作：函数开头就调用await interaction.response.defer()，先告诉Discord“正在处理，稍后返回结果”，抢占3秒时间窗口。
2. 修复响应内容格式化：原代码把字典格式的response直接塞到Embed的description里会显示异常，需要提取choices[0].text字段并做格式处理。
3. 删除无用sleep：await asyncio.sleep(delay=0)完全没必要，直接删掉即可。

修复后的代码：

@bot.tree.command(description="create a txt with chat-gpt")
async def gpt(interaction: discord.Interaction, *, prompt: str):
    # 第一步：立即延迟响应，避免超时
    await interaction.response.defer()
    
    async with aiohttp.ClientSession() as session:
        payload = {
            "model": "text-davinci-003",
            "prompt": prompt,
            "temperature": 0.5,
            "max_tokens": 50,
            "presence_penalty": 0,
            "frequency_penalty": 0,
            "best_of": 1,
        }
        headers = {"Authorization": f"Bearer {API_KEY}"}
        
        async with session.post("https://api.openai.com/v1/completions", json=payload, headers=headers) as resp:
            response = await resp.json()
            
            # 提取并格式化OpenAI的响应内容
            if "choices" in response and len(response["choices"]) > 0:
                gpt_response = response["choices"][0]["text"].strip()
            else:
                gpt_response = "请求失败，请检查API密钥或网络连接。"
            
            # 创建Embed并发送后续响应
            em = discord.Embed(title="Chat GPT的响应:", description=gpt_response)
            await interaction.followup.send(embed=em)

额外注意事项 #

• 确保OpenAI API密钥有效且有足够调用额度。
• 如果OpenAI API调用耗时较长，可以在defer时添加ephemeral=True，让后续响应仅对命令发起者可见，避免其他用户等待。

内容的提问来源于stack exchange，提问作者Pandaroux 1400