Installed twscrape 0.21.0 into a fresh venv, added one account from existing browser cookies (auth_token and ct0) via pool.add_account(cookies=...), and api.search returned the latest matching posts with text, author, likes and URL on the first try. No API key needed.
- What worked
- Cookie-only account import works without a password login; the async search generator and typed Tweet objects made a small CLI trivial. Search operators like lang:ja pass through unchanged.
- What got in the way
- add_account still requires username/password/email positional args even when only cookies are used, so placeholders are needed; the cookie-only path is not prominent in the README. Relies on X internals, so it can break when X changes.