666ghj/MiroFish · critical · ValueError

缺少 API Key 配置，请在项目根目录 .env 文件中设置 LLM_API_KEY

Error message

缺少 API Key 配置，请在项目根目录 .env 文件中设置 LLM_API_KEY

What it means

Raised in backend/scripts/run_parallel_simulation.py during LLM setup: the script reads LLM_API_KEY, copies it into os.environ["OPENAI_API_KEY"] for camel-ai's OpenAI backend, and then raises this ValueError if os.environ.get("OPENAI_API_KEY") is still falsy — i.e. neither LLM_API_KEY nor a pre-existing OPENAI_API_KEY is available. It fails before ModelFactory.create, since every simulation turn would fail authentication.

Source

Thrown at backend/scripts/run_parallel_simulation.py:1027

        llm_model = boost_model or os.environ.get("LLM_MODEL_NAME", "")
        config_label = "[加速LLM]"
    else:
        # 使用通用配置
        llm_api_key = os.environ.get("LLM_API_KEY", "")
        llm_base_url = os.environ.get("LLM_BASE_URL", "")
        llm_model = os.environ.get("LLM_MODEL_NAME", "")
        config_label = "[通用LLM]"
    
    # 如果 .env 中没有模型名，则使用 config 作为备用
    if not llm_model:
        llm_model = config.get("llm_model", "gpt-4o-mini")
    
    # 设置 camel-ai 所需的环境变量
    if llm_api_key:
        os.environ["OPENAI_API_KEY"] = llm_api_key
    
    if not os.environ.get("OPENAI_API_KEY"):
        raise ValueError("缺少 API Key 配置，请在项目根目录 .env 文件中设置 LLM_API_KEY")
    
    if llm_base_url:
        os.environ["OPENAI_API_BASE_URL"] = llm_base_url
    
    print(f"{config_label} model={llm_model}, base_url={llm_base_url[:40] if llm_base_url else '默认'}...")
    
    return ModelFactory.create(
        model_platform=ModelPlatformType.OPENAI,
        model_type=llm_model,
    )


def get_active_agents_for_round(
    env,
    config: Dict[str, Any],
    current_hour: int,
    round_num: int
) -> List:

View on GitHub (pinned to b5b53acc57)

Solutions

1. Add LLM_API_KEY=... to the backend .env (project root) or export it before running.
2. If OPENAI_API_KEY is already exported in the shell, that also passes the check — use it as a fallback.
3. Verify with a preflight: python -c "import os;print(bool(os.environ.get('LLM_API_KEY') or os.environ.get('OPENAI_API_KEY')))".

Example fix

# before
$ python scripts/run_parallel_simulation.py  # no LLM_API_KEY

# after
# .env
LLM_API_KEY=sk-...
LLM_MODEL_NAME=gpt-4o-mini
$ python scripts/run_parallel_simulation.py

import os

def llm_env_ready() -> bool:
    return bool(os.environ.get("LLM_API_KEY") or os.environ.get("OPENAI_API_KEY"))

When it happens

Trigger: Running run_parallel_simulation.py with LLM_API_KEY unset or empty and no pre-exported OPENAI_API_KEY: missing .env, .env not loaded, wrong working directory, or LLM_API_KEY="".

Common situations: New clone without .env; cron/CI shell where .env is not sourced; key stored under a different variable name; .env present but script launched from another directory.

Related errors

• 缺少 API Key 配置，请在项目根目录 .env 文件中设置 LLM_API_KEY
• 缺少 API Key 配置，请在项目根目录 .env 文件中设置 LLM_API_KEY
• ZEP_API_KEY 未配置
• ZEP_API_KEY must be supplied through the process environment
• LLM_API_KEY 未配置

AI-assisted analysis of 666ghj/MiroFish@b5b53acc57 (2026-08-14). Data as JSON: /api/errors/b2de96cf33bbd14d. Report an issue: GitHub.