HKUDS/Vibe-Trading · critical · RuntimeError

langchain-openai is not installed

Error message

langchain-openai is not installed

What it means

The generic OpenAI-compatible path in build_llm needs ChatOpenAI from langchain-openai, but the module import failed (symbol is None). This is the core dependency of the library's LLM stack, so no fallback exists.

Source

Thrown at agent/src/providers/llm.py:1458

        )

    if provider == "deepseek":
        adapter_mode = _deepseek_adapter_mode()
        if adapter_mode != "openai-compatible":
            native_llm = _build_native_deepseek(
                model=name,
                temperature=temperature,
                callbacks=callbacks,
            )
            if native_llm is not None:
                return native_llm
            if adapter_mode == "native":
                raise RuntimeError(
                    "VIBE_TRADING_DEEPSEEK_ADAPTER=native requires langchain-deepseek"
                )

    if ChatOpenAI is None:
        raise RuntimeError("langchain-openai is not installed")
    # MiniMax requires temperature in (0.0, 1.0] — clamp to 0.01 when the
    # default 0.0 is used to avoid an API validation error.
    if provider == "minimax" and temperature <= 0.0:
        temperature = 0.01
    # Kimi reasoning models reject any temperature other than 1
    # ("invalid temperature: only 1 is allowed for this model").
    if (
        caps.name in {"moonshot", "kimi-coding"}
        and _KIMI_FORCED_TEMPERATURE_RE.match(name)
        and temperature != 1.0
    ):
        logger.info("Forcing temperature=1.0 for %s (provider requirement)", name)
        temperature = 1.0
    effort = get_env_config().llm.langchain_reasoning_effort.strip().lower()
    # Moonshot/DeepSeek official APIs emit reasoning by default and ignore this field.
    configured_responses_api = get_env_config().llm.langchain_use_responses_api
    use_responses_api = uses_responses_api(provider, configured_responses_api)
    creds = get_llm_credentials(provider, name)

View on GitHub (pinned to 80ffdda44c)

Solutions

1. Install/repair: pip install langchain-openai
2. If already installed, diagnose the underlying import failure: python -c "from langchain_openai import ChatOpenAI" and fix the reported version conflict (often reinstall langchain-core to matching version)
3. Recreate the virtualenv if dependency resolution is beyond repair

Example fix

# before
RuntimeError: langchain-openai is not installed

# after
pip install -U langchain-openai langchain-core

try:
    from langchain_openai import ChatOpenAI  # noqa: F401
except ImportError as e:
    raise SystemExit(f'Dependency check failed: {e}')

try:
    llm = build_llm()
except RuntimeError as e:
    if 'langchain-openai' in str(e):
        logging.exception('install langchain-openai and restart'); raise

When it happens

Trigger: Any build_llm call for an OpenAI-compatible provider (openai, minimax, kimi, deepseek fallback, etc.) when langchain-openai is missing or its import raised.

Common situations: Broken virtualenv after dependency upgrades; langchain/openai package version conflict causing ImportError at module load; slim production image built without the agent extras.

Related errors

• Anthropic provider requires langchain-anthropic. Install the
• VIBE_TRADING_DEEPSEEK_ADAPTER=native requires langchain-deep
• oauth-cli-kit is not installed. Run: pip install oauth-cli-k
• OpenAI Codex OAuth requires httpx. Install dependencies firs
• {source} produces a non-ASCII HTTP header ({name!r}). Use an

AI-assisted analysis of HKUDS/Vibe-Trading@80ffdda44c (2026-08-28). Data as JSON: /api/errors/c262474f9ab16a07. Report an issue: GitHub.