BerriAI/litellm · error · ImportError

Missing botocore to use AWS SigV4 authentication. Run 'pip i

Error message

Missing botocore to use AWS SigV4 authentication. Run 'pip install boto3'.

What it means

The experimental LiteLLM MCP client's AWS SigV4 auth class needs botocore to build SigV4-signed requests (Credentials, STS assume-role). The import is optional: if neither boto3 nor botocore is installed in the environment, __init__ re-raises ImportError with this message before any connection attempt.

Source

Thrown at litellm/experimental_mcp_client/client.py:122

    for every outgoing request, enabling per-request signature computation.
    """

    requires_request_body = True

    def __init__(
        self,
        aws_access_key_id: str | None = None,
        aws_secret_access_key: str | None = None,
        aws_session_token: str | None = None,
        aws_region_name: str | None = None,
        aws_service_name: str | None = None,
        aws_role_name: str | None = None,
        aws_session_name: str | None = None,
    ):
        try:
            from botocore.credentials import Credentials
        except ImportError:
            raise ImportError("Missing botocore to use AWS SigV4 authentication. Run 'pip install boto3'.")
        self.service_name = aws_service_name or "bedrock-agentcore"
        self.region_name = aws_region_name or "us-east-1"
        # Note: os.environ/ prefixed values are already resolved by
        # ProxyConfig._check_for_os_environ_vars() at config load time.
        # Values arrive here as plain strings.
        if aws_role_name:
            self.credentials = self._assume_role(
                aws_role_name=aws_role_name,
                aws_session_name=aws_session_name,
                aws_access_key_id=aws_access_key_id,
                aws_secret_access_key=aws_secret_access_key,
                aws_session_token=aws_session_token,
                aws_region_name=self.region_name,
            )
        elif aws_access_key_id and aws_secret_access_key:
            self.credentials = Credentials(
                access_key=aws_access_key_id,
                secret_key=aws_secret_access_key,

View on GitHub (pinned to 6c2dcb801b)

Solutions

1. pip install boto3 (pulls botocore) in the environment running the litellm proxy / MCP client
2. Alternatively install just botocore: pip install botocore, if the full boto3 distribution is unwanted
3. Rebuild your Docker image with the extra dependency and restart the proxy
4. If AWS SigV4 auth is not actually needed, remove the aws_* fields from the MCP server config so the SigV4 path is never constructed

Example fix

# before
pip install litellm
# ... AWS SigV4 MCP auth fails with ImportError

# after
pip install litellm boto3

def has_botocore() -> bool:
    try:
        import botocore.credentials  # noqa: F401
        return True
    except ImportError:
        return False

if mcp_auth_uses_sigv4 and not has_botocore():
    raise RuntimeError("pip install boto3 before enabling aws_sigv4_auth")

try:
        from litellm.experimental_mcp_client import MCPClient
    except ImportError as e:
        if "botocore" in str(e): subprocess.run([sys.executable, '-m', 'pip', 'install', 'boto3'])

When it happens

Trigger: Configuring an MCP server entry with aws_sigv4_auth (aws_access_key_id/aws_secret_access_key, or aws_role_name for assume-role) in a venv where 'pip install botocore' or 'pip install boto3' was never run — e.g. a slim Docker image or a fresh proxy deployment that only installed litellm.

Common situations: Running litellm proxy in a minimal container (python:slim) with only litellm installed; enabling Bedrock AgentCore MCP tools after an env that predates boto3; CI environments trimmed of AWS SDK packages.

Understand the failure class

• Authentication and authorization failures — expired tokens, bad credentials, and missing scopes.

Related errors

• No AWS credentials found. Provide aws_access_key_id and aws_
• streamable_http_client is not available. Please install mcp
• Guardrail {self.guardrail_name or type(self).__name__} imple
• Missing boto3 to call bedrock. Run 'pip install boto3'.
• Missing boto3 to call bedrock. Run 'pip install boto3'.

AI-assisted analysis of BerriAI/litellm@6c2dcb801b (2026-08-15). Data as JSON: /api/errors/d524df11f625a571. Report an issue: GitHub.