langchain-ai/langchain · error · NotImplementedError

Unsupported operand type for +: {type(other)}

Error message

Unsupported operand type for +: {type(other)}

What it means

Raised by `PromptTemplate.__add__` when the right operand is neither a `PromptTemplate` nor a `str`. LangChain only knows how to concatenate those two types, so anything else (an int, list, another prompt class like `ChatPromptTemplate`, a `Runnable`) triggers `NotImplementedError`. Note the message interpolates the operand's type, e.g. `Unsupported operand type for +: <class 'int'>`.

Source

Thrown at libs/core/langchain_core/prompts/prompt.py:184

                if k in partial_variables:
                    msg = "Cannot have same variable partialed twice."
                    raise ValueError(msg)
                partial_variables[k] = v
            return PromptTemplate(
                template=template,
                input_variables=input_variables,
                partial_variables=partial_variables,
                template_format=self.template_format,
                validate_template=validate_template,
            )
        if isinstance(other, str):
            prompt = PromptTemplate.from_template(
                other,
                template_format=self.template_format,
            )
            return self + prompt
        msg = f"Unsupported operand type for +: {type(other)}"
        raise NotImplementedError(msg)

    @property
    def _prompt_type(self) -> str:
        """Return the prompt type key."""
        return "prompt"

    def format(self, **kwargs: Any) -> str:
        """Format the prompt with the inputs.

        Args:
            **kwargs: Any arguments to be passed to the prompt template.

        Returns:
            A formatted string.
        """
        kwargs = self._merge_partial_and_user_variables(**kwargs)
        return DEFAULT_FORMATTER_MAPPING[self.template_format](self.template, **kwargs)

View on GitHub (pinned to e32fa9a52e)

Solutions

1. Convert the right operand to a `PromptTemplate` first: `prompt + PromptTemplate.from_template(str(other))`
2. For chat prompts, use `ChatPromptTemplate.from_messages([...])` instead of `+` on a `PromptTemplate`
3. For plain strings no conversion is needed — check that the value is actually a `str` and not `None` (a stray `None` return is a frequent cause)

Example fix

# before
prompt = PromptTemplate.from_template("Q: {question}\n")
extra = some_fn()          # returns None or a non-str object
combined = prompt + extra  # NotImplementedError

# after
extra = some_fn() or ""
assert isinstance(extra, str)
combined = prompt + extra  # str path is supported

def add_to_prompt(prompt: PromptTemplate, other: object) -> PromptTemplate:
    if isinstance(other, PromptTemplate):
        return prompt + other
    if isinstance(other, str):
        return prompt + other
    msg = f"cannot concatenate {type(other).__name__} to PromptTemplate"
    raise TypeError(msg)

from langchain_core.prompts import PromptTemplate

def is_addable(other: object) -> bool:
    return isinstance(other, (PromptTemplate, str)) and other is not None

try:
    combined = prompt + other
except NotImplementedError:
    combined = prompt + PromptTemplate.from_template(str(other))

When it happens

Trigger: `prompt + 5`, `prompt + ["a", "b"]`, or `prompt + chat_prompt` where `chat_prompt` is a `ChatPromptTemplate` (which is a `RunnableSequence`, not a `PromptTemplate`). The isinstance checks for `PromptTemplate` and `str` both fail before the raise.

Common situations: Trying to append a chat-style prompt or a `RunnableLambda` to a string prompt with `+`; assuming all prompt classes are mutually addable; concatenating a formatted string that is actually `None` (e.g. the result of a function that forgot to return).

Related errors

• Cannot add templates of different formats
• Unsupported operand type for +: {type(other)}
• Cannot have same variable partialed twice.
• Structured prompts need to be piped to a language model.
• AsyncTextProjection received a non-string final value

AI-assisted analysis of langchain-ai/langchain@e32fa9a52e (2026-08-14). Data as JSON: /api/errors/604742ccf82b42f6. Report an issue: GitHub.