langchain-ai/deepagents · error · ImportError

Local voice transcription dependencies are missing. Install

Error message

Local voice transcription dependencies are missing. Install the `media` extra and ensure ffmpeg is on PATH.

What it means

Local voice transcription uses Hugging Face `transformers` (loaded lazily via importlib) to build an automatic-speech-recognition pipeline. If `transformers` is not installed, `_load_local_pipeline` raises this ImportError pointing to the `media` extra and ffmpeg. It is a deliberate, explanatory error for an optional dependency set.

Source

Thrown at libs/talon/deepagents_talon/speech.py:253

                wav_path.unlink(missing_ok=True)
            except OSError:
                logger.debug("Could not delete temporary voice transcription file: %s", wav_path)


def _load_local_pipeline(model: str, device: str) -> _LocalSpeechPipeline:
    key = (model, device)
    cached = _local_pipelines.get(key)
    if cached is not None:
        return cached

    try:
        module = importlib.import_module("transformers")
    except ImportError as exc:
        msg = (
            "Local voice transcription dependencies are missing. Install the `media` "
            "extra and ensure ffmpeg is on PATH."
        )
        raise ImportError(msg) from exc

    logger.info("Loading local voice transcription model %s on device=%s", model, device)
    loaded = module.pipeline(
        "automatic-speech-recognition",
        model=model,
        device=device,
    )
    _local_pipelines[key] = loaded
    logger.info("Local voice transcription model %s ready on device=%s", model, device)
    return loaded


def _pipeline_text(result: object) -> str:
    if isinstance(result, Mapping):
        values = cast("Mapping[str, object]", result)
        text = values.get("text")
        return text.strip() if isinstance(text, str) else ""
    text = getattr(result, "text", None)

View on GitHub (pinned to a1af029e6e)

Solutions

1. Install the extra: `pip install "deepagents-talon[media]"` (or `uv sync --extra media`)
2. Verify with `python -c "import transformers"` that the import works in the target environment
3. Alternatively configure a cloud/API transcription backend that does not need the local pipeline

Example fix

# before
pip install deepagents-talon
# after
pip install "deepagents-talon[media]"

import importlib.util
if importlib.util.find_spec("transformers") is None:
    raise SystemExit("Install the media extra: pip install 'deepagents-talon[media]'")

try:
    text = transcribe_local(path)
except ImportError as exc:
    if "media" in str(exc):
        text = transcribe_cloud(path)  # fallback backend
    else:
        raise

When it happens

Trigger: Requesting local (non-API) transcription of an audio file when the `transformers` package (part of the `media` extra) is absent from the environment.

Common situations: Installing talon without `pip install deepagents-talon[media]`; running in a slim Docker image that trimmed ML dependencies; switching from a cloud transcription backend to local transcription in a fresh venv.

Understand the failure class

Background: "X is not installed. Please install it with pip install Y": missing optional dependency errors — ImportError/ValueError raised when a library's optional extra was never installed — this error's family across 22 libraries.

Related errors

• Invalid class_path '{class_path}' for provider '{provider}':
• Could not import module '{module_path}' for provider '{provi
• Class '{class_name}' not found in module '{module_path}' for
• Provider package '{package}' is installed but failed to impo
• Entry point {entry.name!r} does not resolve to a Python modu

AI-assisted analysis of langchain-ai/deepagents@a1af029e6e (2026-08-29). Data as JSON: /api/errors/17f22f9191b886d1. Report an issue: GitHub.