castmd

Cast any page. Feed any model. castmd turns the page you are reading into clean Markdown, ready to paste into ChatGPT, Claude, Gemini, or any other AI tool. One click, clipboard ready. No account, no server, no tracking. Everything runs locally in your browser. WHY MARKDOWN Pasting a URL or raw HTML into an LLM wastes tokens on menus, cookie banners, sidebars and footers, and often confuses the model. castmd strips all of that and keeps what matters: headings, lists, tables, links, code blocks, bold and italic. Cleaner input, fewer tokens, better answers. ONE-CLICK ACTIONS - Copy as Markdown: the whole page, cleaned, on your clipboard - Save as .md: download the page as a Markdown file - Copy Outline: headings only, for mapping a long document fast - Copy all tabs: every open tab in the window merged into one session document - Save all tabs: every open tab written as its own file - Preview and edit: adjust the output before you copy or save it - Right-click any page: "Copy page as Markdown" - Keyboard shortcut: Ctrl+Shift+M (Cmd+Shift+M on Mac), remappable at chrome://extensions/shortcuts BUILT FOR AI WORKFLOWS - Token estimate after every conversion, so you know the cost before you paste - Model fit chips: see at a glance whether the page fits gpt-4, gpt-4o, claude or gemini context windows - Three output modes: Markdown, JSON (url, title, markdown, tokens, timestamp), or XML wrapped in Claude's <document> tags SMART EXTRACTION castmd finds the main content automatically and drops navigation, sidebars and footers. It knows the common article containers (schema.org articleBody, GitHub readme bodies, editorial site wrappers) and falls back to text-density scoring when a page uses none of them. Inline formatting survives the conversion. LOCAL MARKDOWN VIEWER AND EDITOR Open any .md file from your disk and castmd renders it as formatted HTML, GitHub style, in light or dark to match your system. Headings, tables, task lists, fenced code, blockquotes and images all render. Click Edit to change the file in a plain text area, then Save to write it back to disk. YAML frontmatter is hidden in the rendered view and preserved when you save. Turn on "Allow access to file URLs" on the castmd card at chrome://extensions to enable this. CONFLUENCE TREE EXPORT On a Confluence Cloud page, castmd adds a bulk export: the current page plus its descendants as a ZIP of Markdown files that mirrors the page hierarchy. Pick the depth (this page, plus children, or all descendants), preview the page count, then download. Access is requested on demand for your own Confluence domain only, pages you cannot read are skipped and listed in _skipped.txt, and requests are rate limited so your site is not hammered. PRIVACY castmd collects nothing. No analytics, no accounts, no external servers, no remote code. Conversion happens on your machine. The only network requests it ever makes are the Confluence exports you start yourself, sent straight from your browser to your own Confluence site. WHAT CASTMD DOES NOT DO - It does not run on pages you have not acted on - It does not read or send your browsing history anywhere - It does not inject ads, affiliate links or scripts into pages - It does not require sign-in Free and open source. Issues and pull requests welcome. Homepage: https://castmd.cc4.marketing Source: https://github.com/blacklogos/castmd Issues: https://github.com/blacklogos/castmd/issues