mypy error: Need type annotation for (tested)

mypy says Need type annotation for "x" when you assign an empty [], {{}} or set() and nothing in the same scope tells it the element type. Annotate the variable: items: list[str] = [], counts: dict[str, int] = {{}}.

1. Reproduce it, then fix it

b1.py:

items = []
counts = {}
seen = set()
result = None
result = 5

b2.py, the annotated version:

from typing import Optional

items: list[str] = []
counts: dict[str, int] = {}
seen: set[int] = set()
result: Optional[int] = None
result = 5

def collect() -> list[str]:
    out = []          # inferred from the append below: no error
    out.append("a")
    return out

Output (mypy 2.4.0, Python 3.13.5, --no-incremental):

$ mypy b1.py
b1.py:1: error: Need type annotation for "items" (hint: "items: list[<type>] = ...")  [var-annotated]
b1.py:2: error: Need type annotation for "counts" (hint: "counts: dict[<type>, <type>] = ...")  [var-annotated]
b1.py:3: error: Need type annotation for "seen" (hint: "seen: set[<type>] = ...")  [var-annotated]
Found 3 errors in 1 file (checked 1 source file)

$ mypy b2.py
Success: no issues found in 1 source file

• mypy cannot infer an element type from an empty [], {} or set() when nothing later in the same scope adds to it, so it asks for an annotation (error code var-annotated). The hint shows the exact syntax: items: list[str] = [].
• result = None followed by result = 5 was not an error here: mypy accepted it, using the later assignment in the same scope to complete the type. Annotating Optional[int] (or int | None) is still clearer.
• Inside a function, an empty list followed by .append("a") in the same scope is inferred as list[str] without an annotation (collect() in b2.py).

To silence just this check for a line, add # type: ignore[var-annotated]; prefer the annotation.

Written with AI assistance. Measured on 2026-10-23 with mypy 2.4.0 and Python 3.13.5 on Linux. Files were checked verbatim and the output pasted unedited. PlainTables is our extension; mypy is free.