CiteKey: one-click BibTeX with Scholar-matching keys

Generate a complete BibTeX entry for the paper you are reading — one click, without leaving the page. ONE CLICK, NOT FOUR Getting a BibTeX entry from Google Scholar means searching for the paper you already have open, clicking Cite, clicking BibTeX, waiting for a second page, and copying from it. You leave the article to fetch a citation for the article. CiteKey reads the citation metadata that journal pages already embed for indexing services, so the entry is finished before you see it. Click the toolbar button, click Copy. Done, on the page you were already on. IT WORKS WHEN SCHOLAR WILL NOT Scholar throttles traffic that looks automated, and a commercial VPN is often enough to trigger it: thousands of people share those exit addresses, so the traffic from one looks like a bot to Scholar and you get a CAPTCHA instead of a BibTeX link. If you keep a VPN running for other work, that is exactly when Scholar stops cooperating. CiteKey never contacts Scholar — it reads the page already open in front of you — so a VPN changes nothing. KEYS THAT MATCH, SO DUPLICATES GET CAUGHT Reference managers deduplicate on the entry key. Take a paper from Scholar today, then the same paper from the publisher's site next month with a different tool, and you quietly end up with two entries for one paper and two keys in your manuscript. CiteKey builds the key the way Scholar does: first author's surname, the year, and the first significant word of the title, lowercased and reduced to plain ASCII — lecun2015deep, van2008visualizing. Import from both sources and the second collides with the first, so your reference manager flags it instead of duplicating it. READS PDFS, NOT JUST PAGES Often the only copy you can get at is the PDF itself, which carries no citation tags for a tool to read. CiteKey parses the file: the metadata a publisher embeds in it where there is any, and otherwise the layout of the first pages — the title, the authors beneath it, and the venue from the boilerplate. It happens in your browser; the file is never uploaded. WHAT IT HANDLES • Journal articles, conference papers, book chapters, theses, technical reports and preprints, each given the right entry type automatically • Web pages as @online, with url and urldate, and organisations as authors written the way BibTeX needs them • Code repositories, cited by their owner rather than by the page title • Any page carrying standard citation metadata, which covers most academic publishers and institutional repositories AND • Every field editable before you copy, including the citation key • Two layouts: a compact field order, or a spaced layout for web sources • A local library of everything you have copied, searchable, exportable as a single .bib file • Optional lookup in Crossref and OpenAlex when a document's own details are thin or wrong — it asks before it ever touches the network, and shows you what it found, and what that record contradicts, before changing anything • Preprints cited as preprints, with their eprint identifier, rather than as the version they later became • Light, dark and system themes • Keyboard shortcut: Alt+B PRIVACY No account, no analytics, no server, no tracking. CiteKey reads a page — or a PDF — only when you click its button, parses it in your browser, keeps your entries on your own device, and makes no network request at all unless you explicitly ask it to look a paper up. Free and open source: https://github.com/aritraroy24/citekey