BerriAI/litellm · error · ValueError

Unsupported tool type: {tool['type']}

Error message

Unsupported tool type: {tool['type']}

What it means

The tool translation loop matched no known branch: the tool's 'type' is not 'custom'/'function', not a computer_ type, not in ANTHROPIC_HOSTED_TOOLS, not url/mcp/tool_search_*/advisor, so returned_tool and mcp_server are both None and LiteLLM rejects the tool client-side with 'Unsupported tool type: {type}'.

Source

Thrown at litellm/llms/anthropic/chat/transformation.py:771

        elif tool["type"] == ANTHROPIC_ADVISOR_TOOL_TYPE:
            from litellm.types.llms.anthropic import AnthropicAdvisorTool

            _tool_dict: Final = cast(dict, tool)
            advisor_model: Final = _tool_dict.get("model")
            if not isinstance(advisor_model, str):
                raise ValueError("Advisor tool must have a valid model")
            _advisor_tool: Final = AnthropicAdvisorTool(
                type=ANTHROPIC_ADVISOR_TOOL_TYPE,
                name="advisor",
                model=advisor_model,
            )
            if _tool_dict.get("max_uses") is not None:
                _advisor_tool["max_uses"] = _tool_dict["max_uses"]
            if _tool_dict.get("caching") is not None:
                _advisor_tool["caching"] = _tool_dict["caching"]
            returned_tool = _advisor_tool
        if returned_tool is None and mcp_server is None:
            raise ValueError(f"Unsupported tool type: {tool['type']}")

        ## check if cache_control is set in the tool
        _cache_control: Final = tool.get("cache_control", None)
        _cache_control_function: Final = tool.get("function", {}).get("cache_control", None)
        if returned_tool is not None:
            # Only set cache_control on tools that support it (not tool search tools)
            tool_type = returned_tool.get("type", "")
            if tool_type not in (
                "tool_search_tool_regex_20251119",
                "tool_search_tool_bm25_20251119",
            ):
                if _cache_control is not None:
                    returned_tool["cache_control"] = _cache_control
                elif _cache_control_function is not None and isinstance(_cache_control_function, dict):
                    returned_tool["cache_control"] = ChatCompletionCachedContent(**_cache_control_function)

        ## check if defer_loading is set in the tool
        _defer_loading: Final = tool.get("defer_loading", None)

View on GitHub (pinned to 6c2dcb801b)

Solutions

1. Check the exact type string against Anthropic's current docs (dated suffixes must match exactly, e.g. computer_20250124).
2. Upgrade LiteLLM to a version supporting the new tool type.
3. If the tool is a plain function tool, use type 'function' with a 'function' payload.
4. Filter per-provider tool lists so foreign tool types never reach the Anthropic transformation.

Example fix

# before
tools = [{"type": "computer_2024102", ...}]  # typo'd date

# after
tools = [{"type": "computer_20250124", "function": {"name": "computer", "parameters": {...}}}]

SUPPORTED_PREFIXES = ("computer_",)  # plus hosted types from litellm constants
from litellm.constants import ANTHROPIC_HOSTED_TOOLS  # if available

def validate_tool_types(tools, allowed):
    for t in tools:
        typ = t.get("type", "")
        known = typ in ("function", "custom", "url", "mcp",
                        "tool_search_tool_regex_20251119",
                        "tool_search_tool_bm25_20251119") \
                or any(typ.startswith(p) for p in allowed)
        if not known:
            raise ValueError(f"unsupported tool type {typ!r} for anthropic")
    return tools

def tool_type_supported(tool) -> bool:
    typ = tool.get("type", "")
    return (
        typ in ("function", "custom", "url", "mcp",
                "tool_search_tool_regex_20251119", "tool_search_tool_bm25_20251119")
        or typ.startswith("computer_")
    )

When it happens

Trigger: Passing a tool with a typo'd or version-mismatched type string (e.g. 'computer_2024102' or a newer Anthropic type than this LiteLLM version knows), or a provider-specific type from another vendor (e.g. Google function declarations) in an anthropic/ call.

Common situations: LiteLLM version lag after Anthropic ships new tool types (new computer/web_search versions are dated strings); typos in hand-written type fields; reusing tool lists across providers without filtering.

Related errors

• Missing required parameter: parameters
• Missing required parameter: display_width_px or display_heig
• Missing required parameter: name
• Tool search tool must have a valid name
• Advisor tool must have a valid model

AI-assisted analysis of BerriAI/litellm@6c2dcb801b (2026-08-15). Data as JSON: /api/errors/5bdcca8f7b78b38c. Report an issue: GitHub.