QuickstartMoving from RabbitholesCore concepts
Canvas OverviewCreating NodesConnectionsOrganizing the CanvasSearchSide ViewExportKeyboard Shortcuts
Chat NodeText NodeWeb NodePost Node (YouTube & more)Document Node (Beta)Image NodeGroup Node
Models & ProvidersRAG ModePersonasSlash Commands (Skills & Prompts)Using MCP Tools in ChatSub-AgentsVoice Input
Cloud ProvidersAgent ProvidersOllama (Local Models)Custom ProvidersMCP Servers (Beta)Default ModelsConnected ToolsPersona PromptsAccount ManagementSystem Preferences
Troubleshooting & Support

Document Node (Beta)

Upload PDFs, Office files, and more — the node indexes them so you can chat about their content.

Last updated: July 6, 2026

The Document Node brings your files onto the canvas. Drop in a PDF, spreadsheet, or presentation and it's uploaded, converted, and indexed so connected chats can read it.

Also referred to as the "File Node" — its type ID and this page's URL are unchanged.


Supported files

  • Documents: pdf, doc, docx, rtf, txt, md, html
  • Presentations: ppt, pptx
  • Data: xls, xlsx, csv, json, xml

Adding a document

You can add a document three ways:

  1. Press D (or pick the file icon from the toolbar) and choose a file
  2. Right-click the canvas and use the Add menu
  3. Drag and drop a file directly onto the canvas

The file uploads and starts indexing automatically as soon as it's added.


Processing states

The node shows its progress:

  • Starting... — upload is beginning
  • A progress indicator while the file is processed and indexed
  • An error state with a Retry button if something fails
  • Ready — processing is complete, with buttons to open the converted .md and the .pdf

Failed uploads don't retry automatically — click Retry on the node.

PDFs show an inline preview in the node. Non-PDF files show a file-type badge and a Reveal in Folder button to locate the original on disk.


Chatting about the document

Connect the document node to a chat node (or @-mention it from a chat) and its content becomes context.

You can chat about a document immediately — you don't need to wait for processing to finish. Only RAG mode requires processing to be complete, since it depends on the index.

For large documents, RAG mode is the recommended way to chat: it retrieves only the relevant chunks instead of sending the whole file.


Use cases

  • Research — load a folder of papers and use RAG mode to ask questions across all of them without blowing up your token budget.
  • Writing — attach your source material (interviews, reports) and draft with a chat that can cite it.
  • Studying — upload lecture slides (pptx) and readings, then generate summaries and practice questions.
  • Content creation — turn a long PDF report into a thread, script, or newsletter with a connected chat.
  • Planning — drop in spreadsheets (xlsx, csv) of budgets or timelines and ask the AI to spot risks or summarize totals.

Post Node (YouTube & more)

Bring videos and social posts onto the canvas — transcripts and content are fetched automatically.

Image Node

Upload images or generate them with AI, then chat about their content on the canvas.

On this page

Supported filesAdding a documentProcessing statesChatting about the documentUse cases