PairedX
Everything it does

The whole feature set,
at a glance.

A source-linked AI reading workspace — every highlight, note, screenshot, and answer stays pinned to the exact spot it came from. Private, portable, yours.

The PairedX workspace
1
2
3
4
5
1 Library — your documents 2 Tools — highlight, comment, screenshot 3 The paper, with highlights & numbered pins 4 Source-linked notes & AI answers 5 Ask about the whole document
Read & annotate

Highlight text or box a figure.

Select text for a quick Highlight · Note · Ask AI popover, drop a point comment anywhere, or screenshot a figure/equation to ask about it directly.

HighlightCommentScreenshot a regionContinuous or single pageFind in document
The reader toolbar: select, highlight, comment, screenshot tools
Scanned PDFs

Scanned? It reads the page for you — on your device.

Open an image-only PDF and PairedX notices there's no selectable text, then offers one-tap OCR that runs entirely in your browser — your file is never uploaded. It rebuilds a real text layer, so search, highlights, and the AI all work like a normal PDF.

Auto-detectedOn-device — nothing uploadedReal selectable textHighlightable & searchableCached — runs once
PairedX detecting a scanned PDF and offering one-tap on-device OCR
Source-linked

Every note ties back to its exact spot.

A connector line joins each card in the panel to the highlighted passage or captured figure it came from — click a note and the reader jumps right to it.

Connector lineNumbered pinsJump to source
A connector line joining a note to its highlighted passage on the page
Ask AI

Answers pinned to the source.

Ask about any selection; the answer is saved as a note linked to that passage, with rendered math & code and a provenance panel showing exactly what it used.

Source-linked answersLaTeX & code@ai mentionsProvenance chips
An AI answer with rendered content and a provenance panel
Inspectable agent

Watch every step it took.

For deeper questions the AI reads other pages, searches the document, scans the outline, or searches the web — and you can open “the agent's work” to see each call.

Read pageSearch documentOutlineWeb searchCreate visual
The agent trace showing each tool call the AI made
Figures & visuals

Capture a figure — or generate one.

Screenshot any figure or equation and ask about it. The AI can also generate an image or a text/ASCII diagram to explain a concept, grounded in the paper.

Figure captureGenerated imagesASCII diagramsTakeaways
A note built from a captured figure, with an AI answer about it
Notes your way

Show what matters on the card.

Check any message — say the AI's summary instead of your question — to show it in full on the collapsed card. Plus tags, resolve, filters, search, and sort.

Show on cardTagsResolveFilter & searchSort by time / page
A collapsed card showing the full AI answer, tagged AI
Your library

A real multi-document library.

Home, Recents, Starred, and Trash. Star, trash, restore, or delete — the same PDF opened again is recognized by content (SHA-256) so its notes just re-attach.

Home / Recents / Starred / TrashContent-addressedNotes auto-attach
The library sidebar with documents and star/trash actions
Share & export

Send a whole annotated paper.

Share as HTML bundles the PDF + notes into one self-contained file that opens anywhere (read-only) and re-opens in PairedX to keep editing. Or export a clean annotations PDF.

Single-file HTMLRound-trip editingExport to PDFSave / import notes JSON
The Export annotations screen with include options and a live preview
Bring your own model

Your model, your key.

OpenRouter (default) or any OpenAI-compatible endpoint. Your key stays in your browser and is sent per-request — never saved on a server. Toggle visuals and web search.

OpenRouterOpenAI-compatibleBYO keyTool toggles
Settings — AI & Tools: providers, models, and keys
Editable prompts

Tune every prompt.

Edit the system prompts — even the agent's tool descriptions — under Settings → Templates, then export or import your whole set as JSON.

System promptsTool descriptionsExport / importReset to default
Settings — Templates: editable system prompts and tool descriptions
Portable & private

Notes you own, on your machine.

Everything stays in your browser. Optionally sync a portable .notes.json to a folder (Chrome/Edge) so notes travel with your PDFs — great for Drive and other devices.

Browser-only by defaultFolder syncNo cloud uploadNo login
Settings — Storage: browser-only notes with optional folder sync

And everything else

The full list, grouped — every one of these ships today.

Reading

  • Single-page & continuous scroll
  • Zoom, page jump, page thumbnails
  • Find in document (⌘/Ctrl-F)
  • On-device OCR for scanned PDFs
  • Collapse / resize panels

Annotating

  • Highlight text
  • Point comments
  • Screenshot a region
  • Selection popover: Highlight · Note · Ask AI
  • Numbered pins in reading order

AI & agent

  • Ask about a selection or the whole doc
  • Tool-using agent (ReAct)
  • Read page · search · outline · full read
  • Web search · create visual
  • Streaming answers & agent trace
  • Generated images & ASCII diagrams
  • LaTeX, Markdown & code rendering

Notes

  • Compact ↔ expanded cards
  • “Show on card” full-length preview
  • Reply / follow-up in-thread
  • Auto & manual tags
  • Resolve, copy, edit, delete
  • Edit question & re-ask AI

Notes panel

  • Filters: all · unresolved · screenshots · AI · questions
  • Search notes
  • Sort by time or page
  • Auto-scroll to active note

Library

  • Home · Recents · Starred · Trash
  • Star · trash · restore · delete
  • Remove the bundled sample
  • SHA-256 content addressing
  • Storage meter

Open & attach

  • Open a PDF, notes, or shared .html
  • Multi-file & drag-and-drop
  • Notes auto-attach by content
  • “Found notes for this PDF” prompt

Share & export

  • Share as self-contained .html
  • Read-only viewer + made-with banner
  • Re-open shared file to keep editing
  • Export annotations to PDF
  • Save / import notes JSON

Model & prompts

  • OpenRouter or OpenAI-compatible
  • Bring your own key (browser-only)
  • Pick text & image models
  • Editable prompt templates
  • Editable agent tool descriptions
  • Export / import / reset prompts

Storage & privacy

  • Browser-only (IndexedDB + localStorage)
  • Optional folder sync (File System Access)
  • No cloud upload of your PDF
  • No login, no tracking cookies

Your identity

  • Set your name & initials
  • Toggle visuals / web search

Open & free

  • AGPL-3.0, self-hostable
  • Static frontend + serverless AI proxy
  • One-click Vercel deploy

Try it — no signup, no install.

Open the appView on GitHub