json.decoder.JSONDecodeError: Unterminated string starting at: line 1 column 10 (char 9)

The exact error string

The full traceback ends with a line like this:

Traceback (most recent call last):
  File "app.py", line 4, in <module>
    data = json.loads(text)
json.decoder.JSONDecodeError: Unterminated string starting at: line 1 column 10 (char 9)

You may also see it raised through the requests library as requests.exceptions.JSONDecodeError when response.json() parses a body that was cut off mid-string. Both wrap the same json.decoder.JSONDecodeError.

What “starting at: line 1 column 10 (char 9)” means

Unlike Expecting value at char 0, this error means parsing succeeded for a while. The number is the index where the offending string opened — the last " the parser saw before it ran out of characters. To find the problem, slice the input from that index:

text = '{"name": "Alice}'
print(repr(text[9:]))   # '"Alice}' — the value string that was never closed

Cause 1: A missing closing quote

The simplest case — a quote was dropped while hand-editing or hand-building JSON:

import json

text = '{"name": "Alice}'        # closing quote after Alice is missing
data = json.loads(text)
# json.decoder.JSONDecodeError: Unterminated string starting at: line 1 column 10 (char 9)

Cause 2: A backslash escapes the closing quote

A backslash starts an escape sequence in JSON. If a value ends with a backslash, that backslash escapes the closing quote — the parser reads \" as a literal quote character and keeps going, looking for a closing quote that never comes:

import json

# The Python literal '...\\"...' is the JSON text: {"end": "value\"}
text = '{"end": "value\\"}'
data = json.loads(text)
# json.decoder.JSONDecodeError: Unterminated string starting at: line 1 column 9 (char 8)

This usually comes from Windows paths or other values that end in a backslash. The fix is to make sure every backslash in the data is itself escaped — which is exactly what json.dumps() does for you:

import json

# Python value is  C:\Users\  (one trailing backslash)
path = "C:\\Users\\"
text = json.dumps({"path": path})   # dumps escapes it to  {"path": "C:\\Users\\"}
data = json.loads(text)             # round-trips cleanly

Cause 3: Truncated or partial data

If a file was written incompletely, a stream was cut off, or a response was only partially read, the closing quote and braces never arrive. The parser reads to the end and reports the last open string:

import json

# Only the first 20 bytes were read / saved
text = '{"name": "Alexand'
data = json.loads(text)
# json.decoder.JSONDecodeError: Unterminated string starting at: line 1 column 10 (char 9)

print(len(text), repr(text[-30:]))   # confirm the data ends mid-string

Related: a literal newline raises a different error

It is tempting to blame a multiline value, but a raw newline or tab inside a string does not give “Unterminated string.” JSON forbids raw control characters, so the parser stops at the control character itself with a different message:

import json

# '\n' here is a real newline character, not the two characters backslash-n
text = '{"note": "line one\nline two"}'
data = json.loads(text)
# json.decoder.JSONDecodeError: Invalid control character at: line 1 column 19 (char 18)

The two characters backslash-n (written \\n in a Python string, or coming verbatim from a file) are valid JSON and parse fine — it is the real line break that breaks. Escape newlines as \n, or let json.dumps() do it. See Bad control character in string literal in JSON for that error in detail.

How to debug it

Print the length and the tail of the input, then slice from the reported index to isolate the open string:

import json

try:
    data = json.loads(text)
except json.JSONDecodeError as e:
    print(f"{e.msg} at char {e.pos}")     # the message and the start index
    print("length:", len(text))
    print("from error:", repr(text[e.pos:e.pos + 40]))
    print("tail:", repr(text[-40:]))

The fix: let json.dumps() build the JSON

Almost every version of this error comes from JSON that was typed or concatenated by hand. Serialize Python objects instead — json.dumps() quotes and escapes every string, including newlines, tabs, and backslashes:

import json

record = {
    "name": "Alice",
    "note": "line one\nline two",   # newline is fine — dumps() escapes it
    "path": "C:\\Users\\alice",
}

text = json.dumps(record)   # guaranteed valid JSON
data = json.loads(text)     # round-trips without error