deepset-ai/haystack · error · ValueError

Number of replies ({len(replies)}), and metadata ({len(meta)

Error message

Number of replies ({len(replies)}), and metadata ({len(meta)}) must match.

What it means

AnswerBuilder.run raises this when the `replies` and `meta` lists passed in have different lengths. Every reply must have a corresponding metadata entry so the builder can attach per-answer metadata consistently.

Source

Thrown at haystack/components/builders/answer_builder.py:181

                    `[^\\n]+$` finds "this is an answer" in a string "this is an argument.\\nthis is an answer".
                    `Answer: (.*)` finds "this is an answer" in a string
                    "this is an argument. Answer: this is an answer".
        :param reference_pattern:
            The regular expression pattern used for parsing the document references.
            If not specified, no parsing is done, and all documents are returned.
            References need to be specified as indices of the input documents and start at [1].
            Example: `\\[(\\d+)\\]` finds "1" in a string "this is an answer[1]".
        :param expand_reference_ranges:
            If True, reference ranges like `[6-10]` are expanded to documents 6 through 10.
            If not specified, the value from the component initialization is used.

        :returns: A dictionary with the following keys:
            - `answers`: The answers received from the output of the Generator.
        """
        if not meta:
            meta = [{}] * len(replies)
        elif len(replies) != len(meta):
            raise ValueError(f"Number of replies ({len(replies)}), and metadata ({len(meta)}) must match.")

        if pattern:
            AnswerBuilder._check_num_groups_in_regex(pattern)

        pattern = pattern or self.pattern
        reference_pattern = reference_pattern or self.reference_pattern
        expand_reference_ranges = (
            self.expand_reference_ranges if expand_reference_ranges is None else expand_reference_ranges
        )
        reference_pattern = AnswerBuilder._resolve_reference_pattern(
            reference_pattern=reference_pattern, expand_reference_ranges=expand_reference_ranges
        )

        replies_to_iterate = replies[-1:] if self.last_message_only and replies else replies
        meta_to_iterate = meta[-1:] if self.last_message_only and meta else meta

        all_answers = []
        for reply, given_metadata in zip(replies_to_iterate, meta_to_iterate, strict=True):

View on GitHub (pinned to e318778c9b)

Solutions

1. Make len(meta) equal len(replies) — provide one meta dict per reply.
2. If metadata is uniform, pass meta=[meta_dict] * len(replies) (or simply omit meta, which defaults to empty dicts for all replies).
3. Check the upstream generator's number of completions (e.g. n parameter) and build meta accordingly.

Example fix

// before
answer_builder.run(replies=replies, meta=[single_meta])
// after
answer_builder.run(replies=replies, meta=[single_meta] * len(replies))

if meta and len(meta) != len(replies):
    raise AssertionError(f"replies={len(replies)} vs meta={len(meta)}")

try:
    result = answer_builder.run(replies=replies, meta=meta)
except ValueError as e:
    if "must match" in str(e):
        logging.error("Replies/meta mismatch: %s", e)
    raise

When it happens

Trigger: Calling answer_builder.run(replies=[...], meta=[...]) where len(meta) != len(replies); e.g. passing meta from a single reply with a multi-reply generator output (or vice versa).

Common situations: Wiring a generator that returns multiple completions into AnswerBuilder while meta was built for one reply; manually constructing meta with fewer entries; schema drift between generator versions.

Related errors

• Pattern '{pattern}' contains multiple capture groups. Please
• The following tool names are not valid: {invalid_tool_names}
• user_prompt must define exactly one message block, found {le
• system_prompt must define exactly one message block, found {
• LLM evaluator expects all input lists to have the same lengt

AI-assisted analysis of deepset-ai/haystack@e318778c9b (2026-08-30). Data as JSON: /api/errors/207b3adddee2b58f. Report an issue: GitHub.