Skip to main content
This quickstart gets you up and running with Graphor. You’ll upload a document, apply advanced parsing, extract structured data, and chat with your content.

​
Step 1: Create Your Project

  1. Sign up for a Graphor Cloud account
  2. Login at app.graphorlm.com/login
  3. Click New project and give it a name
Project list demo

​
Step 2: Upload a Document

The fastest way to add a document is to drag and drop it directly onto the Sources page. Sources page Graphor supports PDFs, Word documents, images, spreadsheets, and 25+ other formats.
You can also ingest content from URLs, GitHub repositories, and YouTube videos. See the Data Ingestion guide for all options.
After uploading, Graphor automatically processes your document using the Fast parsing method. Wait for the status to change to Processed.
Navigation tip: Single-click on a source to select it. Double-click to open Source details.

​
Step 3: Apply Advanced Parsing

For complex documents, you can apply more powerful parsing methods:
MethodBest For
FastSimple text documents (default)
Hi-ResComplex layouts with OCR
Hi-Res FTSpecialized documents requiring highest accuracy
MAIComplex layouts, manuscripts, handwritten documents
GraphorMulti-page tables, diagrams, images with rich annotations
To apply a different parsing method:
  1. Double-click on your document to open Source details
  2. Go to the Settings tab
  3. Select a parsing method
  4. Click Parse
Parsing settings

​
Viewing Results

After parsing, check the Results tab to see:
  • Document view — Visual preview of parsed content
  • Markdown view — Raw text output
  • Page navigation — Browse multi-page documents
Parsing results Each parsing run is saved in the Versions panel on the left. You can compare different methods and switch the active version at any time.
For detailed information about parsing methods and options, see the Data Ingestion guide.

​
Step 4: Extract Structured Data

Transform your documents into structured data using VLM-powered extraction.
  1. In Source details, go to the Extraction tab
  2. Add instructions to guide the extraction (optional)
  3. Define your schema with fields:
    • Field name: e.g., invoice_number
    • Field type: Text, Number, Boolean, or Array
    • Description: What to extract
  4. Click Extract
Example for invoice extraction: Instructions:
Extract all invoice information. Use YYYY-MM-DD for dates and USD for amounts.
Schema:
Field NameTypeDescription
invoice_numberTextThe unique invoice identifier
dateTextInvoice date in YYYY-MM-DD format
total_amountNumberTotal amount due
Data extraction The results show extracted data with page references indicating where each item was found.
For more schema examples and best practices, see the Data Extraction guide.

​
Step 5: Chat with Your Documents

Click the Chat button in the top-right corner to start asking questions. Two chat options:
  • Sources page → Chat with all documents
  • Source details page → Chat with one specific document
Document chat Ask questions in natural language:
You: What are the main findings in this document? Chat: The document presents three main findings… You: Tell me more about the first one.
The chat maintains conversational memory, so follow-up questions understand previous context.
For tips on getting better answers, see the Document Chat guide.

​
What’s Next?

RAG Quickstart

Build your first RAG pipeline with Dataset, Chunking, Retrieval, and LLM nodes

Data Ingestion

Deep dive into parsing methods, multiple source types, and OCR options

Data Extraction

Advanced schema design and extraction best practices

Document Chat

Tips for asking better questions and API integration