Python: 'json.decoder.JSONDecodeError: Expecting value: line 1 column 1 (char 0)' when calling response.json() or json.loads(). What does it mean and how do I fix it?

Cause: the text you tried to parse is not JSON at all; it failed on the very first character. With HTTP APIs (requests, httpx, LLM or REST APIs) the body is usually empty, an HTML page (login page, error page from a proxy, CDN or rate limiter, a 404 page), or plain text, often because the request itself failed. With files, the file is empty, the wrong file, or starts with a byte-order mark. (In requests 2.27+ the same failure shows as requests.exceptions.JSONDecodeError.) Fix: 1. Look at what you actually got before parsing: print(response.status_code, response.headers.get('Content-Type')) and print(response.text[:500]). 2. Check the status first: call response.raise_for_status() (or check response.ok) before response.json(), so 401/403/404/429/5xx errors show up as HTTP errors instead of JSON errors. 3. Only parse JSON when the server says it's JSON (Content-Type: application/json). If the API needs it, send Accept: application/json and the right auth headers; HTML often means a login redirect or a blocked request. 4. Handle empty bodies: 204 No Content and some successful calls return nothing; check if response.content before parsing. 5. Common URL mistakes: http vs https, a missing /api prefix, or a trailing slash redirect that lands on an HTML page. 6. For files: make sure the path is right and the file isn't empty; open with encoding='utf-8-sig' if it starts with a BOM; for JSON Lines (one object per line) parse each line separately. 7. For LLM outputs you asked to be JSON: the model may add prose or code fences; use the provider's structured output / JSON mode, or strip fences and validate before json.loads(). Verify: the printed body is valid JSON (it starts with { or [), and the call succeeds after the status check. Sources: https://docs.python.org/3/library/json.html https://requests.readthedocs.io/en/latest/user/quickstart/