BerriAI/litellm · error · OCIError

OCI only supports function tool calls

Error message

OCI only supports function tool calls

What it means

The OCI GENERIC inference API only supports function-type tool calls. During adaptation, any tool call whose 'type' field is not exactly 'function' raises OCIError(400) 'OCI only supports function tool calls'. This rejects future/alternate OpenAI tool call kinds (e.g. 'custom_function', 'code_interpreter') before the request is sent.

Source

Thrown at litellm/llms/oci/chat/generic.py:121

                )
            new_content.append(OCIImageContentPart(imageUrl=OCIImageUrl(url=image_url)))

    return OCIMessage(
        role=open_ai_to_generic_oci_role_map[role],
        content=new_content,
        toolCalls=None,
        toolCallId=None,
    )


def adapt_messages_to_generic_oci_standard_tool_call(role: str, tool_calls: list) -> OCIMessage:
    """Convert an assistant tool-call message to OCI format."""
    tool_calls_formatted: Final = []
    for tool_call in tool_calls:
        if not isinstance(tool_call, dict):
            raise OCIError(status_code=400, message="Each tool call must be a dictionary")
        if tool_call.get("type") != "function":
            raise OCIError(status_code=400, message="OCI only supports function tool calls")

        tool_call_id = tool_call.get("id")
        if not isinstance(tool_call_id, str):
            raise OCIError(status_code=400, message="Tool call `id` must be a string")

        tool_function = tool_call.get("function")
        if not isinstance(tool_function, dict):
            raise OCIError(status_code=400, message="Tool call `function` must be a dictionary")

        function_name = tool_function.get("name")
        if not isinstance(function_name, str):
            raise OCIError(status_code=400, message="Tool call `function.name` must be a string")

        arguments = tool_call["function"].get("arguments", "{}")
        if not isinstance(arguments, str):
            raise OCIError(
                status_code=400,
                message="Tool call `function.arguments` must be a JSON string",

View on GitHub (pinned to 6c2dcb801b)

Solutions

1. Set 'type':'function' (exact lowercase) on every tool call you send to OCI.
2. Filter or rewrite non-function tool calls from history before routing the conversation to oci/ models.
3. Keep alternate tool-call kinds on a provider that supports them.

Example fix

# before
{'id':'c1','type':'custom_function','function':{...}}

# after
{'id':'c1','type':'function','function':{'name':'get_weather','arguments':'{}'}}

for msg in messages:
    for tc in msg.get('tool_calls') or []:
        if tc.get('type') != 'function':
            tc['type'] = 'function'  # or drop/reject per your policy

def is_function_tool_call(tc: object) -> bool:
    return isinstance(tc, dict) and tc.get('type') == 'function'

try:
    litellm.completion(model='oci/...', messages=msgs)
except OCIError as e:
    if 'only supports function tool calls' in str(e):
        msgs = [m for m in msgs if keeps_function_calls_only(m)]
    raise

When it happens

Trigger: Sending an assistant history message containing tool_calls with type 'custom_function' or any non-'function' value to an oci/ GENERIC model; round-tripping responses from providers that emit other tool call types.

Common situations: Porting an agent framework that defines custom tool call types; replaying a conversation captured on another provider; typos like 'Function' (case-sensitive comparison) or 'functions'.

Related errors

• Content type `{item_type}` is not supported by OCI
• Each tool call must be a dictionary
• Tool call `id` must be a string
• Tool call `function` must be a dictionary
• Tool call `function.name` must be a string

AI-assisted analysis of BerriAI/litellm@6c2dcb801b (2026-08-15). Data as JSON: /api/errors/a0c2048c40ad593b. Report an issue: GitHub.