Perplexity is the only major AI tool that grounds every response in real-time web sources and numbers every claim. That citation structure is valuable — and it survives the export intact. MarkdownTools preserves the numbered sources and turns the whole response into a professional PDF in 30 seconds.
Unlike ChatGPT or Claude, Perplexity AI retrieves live web sources for every query and attaches numbered citations to each factual claim. The raw output is markdown with inline citation brackets — [1], [2], [3] — and a numbered source list at the bottom. This structure makes Perplexity responses uniquely suited for PDF export: the citation apparatus is already there.
The challenge is preserving those citations through the copy-paste step. Copying rendered text loses the bracket notation. Copying via the copy icon preserves it. Once in MarkdownTools, citation numbers render inline and the source list renders as a clean numbered References section at the bottom of the PDF.
Inline citations
Numbered [1][2][3] on every claim
Source list
Ordered list with URLs at the bottom
Structured sections
H2/H3 headings throughout
Typical Perplexity output
## Key Findings Large language models reached human-level performance on most medical licensing exams by early 2024 [1]. This performance persisted across specialties including radiology and pathology [2][3]. Critics note that exam performance does not equal clinical judgment [4]. ## Sources 1. Smith et al., NEJM 2024 2. Chen et al., Radiology 2024 3. Park et al., Pathology 2024 4. Editorial, Lancet 2024
Four steps. Works for chat responses, Pro research reports, and Perplexity Pages.
In Perplexity, click the copy icon below the response — not Ctrl+A in the text area. The copy icon captures the full markdown including the numbered citation references [1][2][3] inline and the source list at the bottom. Copying by selecting text strips the citation brackets and source URLs. About 5 seconds.
Open MarkdownTools and paste into the editor. The live preview on the right renders the headings, bullet points, and inline citation numbers immediately. The numbered source list at the bottom of a Perplexity response renders as an ordered list — exactly right for a reference section in a PDF. 5 seconds.
Scroll to the bottom of the preview and check the source list. Perplexity formats citations as a numbered list with URLs. If any citation lines look broken, use the editor to add a blank line before the list — this ensures the markdown parser treats it as a standalone ordered list. 10 seconds.
Choose a theme — Classic works well for academic and research documents, Clean for technical reports — then click the PDF button. The PDF generates in under two seconds. Citations appear at the bottom of the document exactly as they do in the preview. Your content is never stored. 5 seconds.
Citations are the thing that makes Perplexity output uniquely valuable. Here is how to keep them intact.
Always use the copy icon, not text selection
The copy icon in Perplexity captures the raw markdown. Text selection captures the rendered output — citation brackets often get stripped or transformed when copying rendered HTML. One extra click saves you the cleanup.
Check for a blank line before the numbered source list
Markdown parsers require a blank line before an ordered list when the preceding element is a paragraph. If your source list runs together with the last paragraph, add a blank line in the MarkdownTools editor. The list renders correctly immediately.
Leave citation numbers inline — do not remove them
The inline [1][2][3] numbers look academic but they are the credibility signal. Readers of a fact-checked report expect to see them. In the PDF they appear as plain text brackets, which is the standard format for inline citations in many academic styles.
For Pro reports with 20+ sources, consider trimming before export
A full source list from a Perplexity Pro deep research session can add one or more pages to the PDF. If you are producing a summary document, remove sources that are not directly cited in the body. Keep the full list for anything that will be reviewed for factual accuracy.
Format the source section heading consistently
Perplexity sometimes labels the source section "Sources" and sometimes leaves it unlabeled. In the editor, add "## Sources" or "## References" above the list if it is missing — this gives the section a proper heading in the PDF and makes it easier to navigate.
Verify URLs in the source list before sharing the PDF
Perplexity sources URLs from the live web. Occasionally a URL is paywalled, redirected, or no longer active. If the PDF will be shared with readers who may follow the citations, quickly verify the key sources render or link correctly before distributing.
Perplexity Pages are structured long-form documents. They need slightly different handling than a chat response.
What Perplexity Pages are
Pages is a Perplexity feature that generates publication-style long-form content: a title, introduction, multiple H2 sections, and conclusion. It is more like a structured article than a chat answer, and the markdown is correspondingly more elaborate — deeper heading nesting, longer paragraphs, and sometimes blockquotes for key findings.
How to copy a Page correctly
Open the Page and use the copy or export option in the Pages editor. Do not use Ctrl+A / Cmd+A from the rendered page view — that captures UI elements (navigation, share buttons) alongside the content. The copy option in the editor gives you the raw markdown document without UI chrome.
What renders well from Pages
The full heading hierarchy of a Perplexity Page — title, H2 sections, H3 subsections — renders correctly in MarkdownTools. Blockquotes used for key findings render with a left border, which looks professional in PDF. Embedded inline citations carry through the same as in chat responses.
What may need cleanup in Pages
Pages sometimes include image placeholders or caption markers that do not have corresponding images in the markdown. These appear as broken image alt text in the preview. Delete them in the editor before exporting — they are typically short lines starting with an exclamation mark.
Choosing a theme for Pages
Pages are content-heavy long-form documents. Classic theme works well — it uses slightly more generous line height and serif-influenced type that makes long-form reading comfortable in PDF. Clean is better for technically structured Pages with many code examples or data tables.
Long Pages and pagination
A Perplexity Page can be several thousand words. MarkdownTools handles this within normal limits. If the PDF has awkward page breaks — a heading at the bottom of a page with the content on the next — this is normal PDF pagination behavior. It does not indicate a problem with the content.
Problems specific to Perplexity output and how to resolve them in under a minute.
Citation numbers [1][2] showing as raw text instead of rendering
Why: This is expected — citation brackets are not special markdown syntax. They render as plain text, which is correct. Some users expect them to be rendered as superscripts or links.
Fix: No fix needed. [1] appearing inline in the PDF body is the standard format for inline citations in many citation styles. If you need hyperlinked citations, manually convert the source URLs in the editor to markdown links.
Source list at the bottom not rendering as a numbered list
Why: The numbered list is missing a blank line before it. When a numbered list immediately follows a paragraph with no blank line between them, some markdown parsers treat it as a continuation of the paragraph.
Fix: In the MarkdownTools editor, place the cursor between the last paragraph and the first numbered source and press Enter to add a blank line. The preview updates immediately and the list renders correctly.
Perplexity response pasted but citation brackets are missing
Why: Text was copied by selecting it rather than using the copy icon. The rendered Perplexity interface transforms citation brackets into styled elements, and text selection copies the rendered output — brackets often get dropped.
Fix: Go back to Perplexity and use the copy icon below the response. This captures the raw markdown including all citation brackets. Re-paste into MarkdownTools.
Long research report breaks pagination badly
Why: A 3,000+ word Perplexity Pro report will have multiple page breaks. Some may fall mid-section or in awkward spots relative to headings.
Fix: Switch to the Clean theme, which uses more compact typography and may shift break points. For reports where pagination matters a great deal, consider splitting at major H2 sections and exporting as separate PDFs.
Image placeholders from Perplexity Pages appear in the PDF
Why: Perplexity Pages sometimes include image references in the markdown. Without actual image files, these render as broken image alt text.
Fix: In the editor, find lines starting with " — or by wrapping the URL in angle brackets. The PDF will display the link text rather than the raw URL.
Anywhere cited AI research needs to leave the browser and become a shareable, printable document.
Perplexity synthesizes information from multiple web sources and numbers every claim. When you export that response as a PDF via MarkdownTools, the numbered citations carry through. Readers can see exactly which sources back which claims — the same academic rigor as a cited report, produced in 30 seconds.
Ask Perplexity to summarize the current state of research on a topic. It will return a structured response with section headings and a numbered source list at the bottom. Paste into MarkdownTools, apply the Classic theme, and download a literature review PDF ready to attach to a grant application, thesis proposal, or internal briefing.
Perplexity Pro users can run deep research sessions that compile competitor data, market positions, and pricing from live web sources. The structured output — with sections like "Market Position," "Key Features," and "Pricing" — converts directly into a scannable competitive analysis PDF with citation numbers linking each claim to its source.
The distinguishing feature of Perplexity output is that every factual claim carries a citation number. When you export this as a PDF, those numbers remain. A reader who wants to verify a specific claim can cross-reference the source list at the bottom. This makes Perplexity-sourced PDFs significantly more credible than AI output from models that do not cite sources.
Perplexity Pages produces long-form, structured content — introductions, multi-section bodies, conclusions — formatted as a document rather than a chat response. Copying the raw content and pasting into MarkdownTools gives you a PDF that mirrors the original page structure: proper heading hierarchy, ordered sections, and any embedded source references.
Students and researchers use Perplexity to build initial literature maps and quick factual foundations before diving into primary sources. Exporting a Perplexity session as a PDF gives you a shareable, printable record of the AI-assisted research phase — useful for supervisors, collaborators, or your own offline review with the source list intact for follow-up verification.
Perplexity AI is built around real-time web search. Every response is grounded in sources retrieved at query time, and every factual claim is attached to a numbered citation. In the raw markdown, those citations appear as superscript-style brackets inline — [1], [2], [3] — and a numbered source list appears at the bottom of the response. This structure is unique among consumer AI tools.
Most other AI outputs are prose with formatting. Perplexity outputs are more like annotated research briefs — prose, formatting, and an embedded citation apparatus. That citation structure is meaningful: stripping it by copying rendered text loses the source attributions. Using the copy icon preserves it.
The practical implication for PDF export: Perplexity markdown is already more document-like than a typical AI chat response. The numbered source list renders as an ordered list in MarkdownTools, which maps naturally to a References or Sources section at the bottom of the PDF. You are not fighting the format — you are working with it.
The most important step in the Perplexity-to-PDF workflow is using the copy icon rather than selecting text. Perplexity renders markdown in the browser — when you see "[1]" in the Perplexity interface it is already rendered as a styled superscript, not raw text. Copying the rendered view via text selection captures the visual output and often drops the bracket notation.
The copy icon captures the underlying markdown. That means "[1]" stays as "[1]" in the clipboard, the numbered source list stays as a markdown ordered list, and all source URLs are preserved. Paste this into MarkdownTools and the citation numbers appear inline exactly where Perplexity placed them, with the source list formatted as a clean ordered list at the bottom.
If you are copying from a long Perplexity response that spans multiple sections, check that the source list at the bottom has a blank line before it. Markdown parsers require a blank line before an ordered list when the preceding line is not also a list item. If the list looks malformed in the preview, add that blank line in the editor — it takes two seconds and fixes the rendering.
Perplexity has two distinct output modes that require slightly different handling. Chat responses are the standard answer-to-a-question format: an opening paragraph, structured body, and source list. These are straightforward to export — copy, paste, export.
Perplexity Pages is a separate feature that generates long-form, publication-style content. Pages have a title, an introduction, multiple H2 sections, and sometimes embedded media references. The markdown structure of a Page is more elaborate than a chat response — deeper heading nesting, longer paragraphs, and sometimes blockquotes for key findings.
For Perplexity Pages, access the raw content via the share or copy option in the Page editor, not by copying from the rendered view. The rendered view of a Page includes UI chrome (share buttons, navigation) that will appear in the clipboard if you select-all. The raw markdown copy gives you just the document content. Paste into MarkdownTools and the full page structure renders correctly — heading hierarchy, section bodies, and any inline citations from the page.
Perplexity Pro users can generate deep research reports — extended, multi-source documents that can run to several thousand words with 20 or more citations. These are among the most structured AI outputs available for research tasks, with section headings, sub-sections, and a comprehensive source list.
MarkdownTools handles documents up to 10,000 lines. A full Perplexity Pro research report typically runs 1,500 to 4,000 words, well within that limit. PDF generation for a report that size completes in under three seconds.
One practical note for long Perplexity reports: the source list can be extensive. Twenty or more numbered entries at the bottom of the document will add roughly one full page to the PDF. If you want a more compact output, you can trim the source list to the most important sources before exporting — or keep it complete if the document will be used in any academic or professional context where full attribution matters.
Everything you need to know.
Use the copy icon below the Perplexity response, not Ctrl+A or Cmd+A to select text. The copy icon captures the raw markdown, which includes the citation brackets [1][2][3] inline and the full numbered source list at the bottom. Selecting text copies the rendered output, which often strips the bracket notation and loses source URLs. Once you paste into MarkdownTools, the citation numbers appear exactly as they were in the Perplexity response.
Paste it. Citations stay intact. Download. No signup, no cost.