langchain-ai/deepagents · error · NotImplementedError

Audio is not currently supported.

Error message

Audio is not currently supported.

What it means

`convert_audio_block_to_content_blocks` in libs/acp/deepagents_acp/utils.py unconditionally raises NotImplementedError: audio prompt blocks cannot yet be translated to LangChain content blocks. The ACP bridge supports text, image, resource, and embedded-resource blocks but not audio.

Source

Thrown at libs/acp/deepagents_acp/utils.py:42

def convert_image_block_to_content_blocks(block: ImageContentBlock) -> list[dict[str, object]]:
    """Convert an ACP image block to LangChain content blocks."""
    # Primary case: inline base64 data (data is already a base64 string)
    if block.data:
        data_uri = f"data:{block.mime_type};base64,{block.data}"
        return [{"type": "image_url", "image_url": {"url": data_uri}}]

    # No data available
    return [{"type": "text", "text": "[Image: no data available]"}]


def convert_audio_block_to_content_blocks(block: AudioContentBlock) -> list[dict[str, str]]:
    """Convert an ACP audio block to LangChain content blocks.

    Raises:
        NotImplementedError: Audio content is not yet supported.
    """
    msg = "Audio is not currently supported."
    raise NotImplementedError(msg)


def convert_resource_block_to_content_blocks(
    block: ResourceContentBlock,
    *,
    root_dir: str,
) -> list[dict[str, str]]:
    """Convert an ACP resource block to LangChain content blocks."""
    file_prefix = "file://"
    resource_text = f"[Resource: {block.name}"
    if block.uri:
        # Truncate root_dir from path while preserving file:// prefix
        uri = block.uri
        has_file_prefix = uri.startswith(file_prefix)
        path = uri[len(file_prefix) :] if has_file_prefix else uri

        # Remove root_dir prefix to get path relative to agent's working directory
        if path.startswith(root_dir):

View on GitHub (pinned to a1af029e6e)

Solutions

1. Remove the audio block from the prompt content before calling `prompt()`; send a text transcription instead
2. Transcribe the audio client-side (e.g. with a speech-to-text model) and pass the text as a TextContentBlock
3. If you control the client, filter unsupported block types before sending
4. Track library updates for audio support before retrying audio blocks

Example fix

# before
prompt_blocks = [AudioContentBlock(...), text_block]
# after
prompt_blocks = [TextContentBlock(type="text", text=transcribe(audio))]

if any(isinstance(b, AudioContentBlock) for b in prompt_blocks):
    raise ValueError("ACP transport does not support audio blocks; transcribe to text first")

def is_supported_block(block: object) -> bool:
    return not isinstance(block, AudioContentBlock)

try:
    resp = await conn.prompt(blocks, session_id)
except NotImplementedError as exc:
    if "Audio is not currently supported" in str(exc):
        blocks = [b for b in blocks if not isinstance(b, AudioContentBlock)]
        resp = await conn.prompt(blocks, session_id)
    else:
        raise

When it happens

Trigger: Sending a `prompt()` request whose content list contains an `AudioContentBlock` (ACP `audio` block with `AudioContentBlockData`) — conversion is attempted in server.py:983-984.

Common situations: An ACP client that records or forwards voice input; a protocol client configured to attach microphone audio; automated tests or proxies forwarding arbitrary block types to the agent.

Related errors

• Agent initialization failed
• -32600
• Could not parse embedded resource block. Block expected eith
• Default backend doesn't support command execution (SandboxBa
• NotImplementedError raised by abstract `ls` (backend does no

AI-assisted analysis of langchain-ai/deepagents@a1af029e6e (2026-08-29). Data as JSON: /api/errors/9fb059fadee590a8. Report an issue: GitHub.