Complete guide to using DocuMind AI
DocuMind AI is an intelligent document assistant that allows you to upload PDF documents and ask questions about them using advanced AI technology. Our system uses Retrieval-Augmented Generation (RAG) to provide accurate, context-aware answers based on your document content.
Visit the signup page and create your account with your email and password.
Click the "Upload PDF" button on your dashboard, select a PDF file, and wait for it to be processed. The system will extract text, create embeddings, and make it searchable.
Navigate to the "Ask Questions" page, select your document, and start asking questions. The AI will search through your document and provide accurate answers.
To get started with DocuMind AI, you need to create an account. This ensures your documents are secure and isolated from other users.
Once you've created your account, here's what you should do:
Currently, DocuMind AI supports PDF files (.pdf) only. We're working on adding support for other document formats in future updates.
Uploading a document is a simple process that happens in several stages:
Click "Upload PDF" and select your file, or drag and drop it into the upload area.
The file is uploaded to secure cloud storage (0-60% progress).
Text is extracted from the PDF and split into manageable chunks (60-90% progress).
AI embeddings are generated for semantic search capabilities.
Document is indexed in the vector database and ready for queries (90-100% progress).
Your dashboard provides comprehensive document management features:
See all your uploaded documents with file size, upload date, and title.
Click the message icon to ask questions about a specific document.
Remove documents and all associated Q&A history with the delete button.
Easily upload additional documents from the dashboard or upload page.
File size limits depend on your subscription plan:
| Plan | Max File Size | Total Storage |
|---|---|---|
| Free | 10 MB | 100 MB |
| Basic | 50 MB | 5 GB |
| Pro | 200 MB | Unlimited |
You can ask questions in natural language. The AI understands various question types:
"What is the contract duration?" or "What is the salary mentioned?"
"Summarize the key terms" or "What are the main points?"
"What are the differences between section A and B?"
"What does clause 5.2 say?" or "Who are the parties involved?"
Specific questions yield better results than vague ones.
Reference specific sections, dates, or terms when possible.
Multiple questions in one query may reduce answer quality.
If the answer isn't satisfactory, try rephrasing or being more specific.
Answers are generated using Retrieval-Augmented Generation (RAG), which means:
Answers preserve line breaks and formatting, making them easy to read. Multi-paragraph answers are displayed with proper spacing.
The chat interface provides a conversational experience:
Your questions appear immediately when submitted, shown in blue bubbles on the right.
While processing, you'll see a "Thinking..." indicator with a spinner.
Answers appear in gray bubbles on the left, preserving formatting and line breaks.
Press Enter to send, Shift+Enter for a new line.
Your dashboard is the central hub for managing your documents and viewing your activity:
Access Pricing, FAQs, Docs, Upload, and your account menu.
View total documents, questions asked, and account status at a glance.
See all your documents with quick actions to ask questions or delete.
Review your recent questions and answers for quick reference.
The dashboard displays three key statistics:
The document list shows:
Your Q&A history displays:
Note: Q&A history is tied to specific documents. When you delete a document, its Q&A history is also removed.
The Account Status page provides comprehensive information about your account:
View your current plan and available features.
DocuMind AI takes security seriously:
Each user's documents are completely isolated. You can only access your own documents.
Documents are stored in secure cloud storage with encryption.
All API requests require user authentication via user ID headers.
We never share your documents or data with third parties.
All API requests require user authentication. You can provide your user ID in two ways:
// Method 1: Via Header (Recommended)
X-User-ID: your-user-id
// Method 2: Via Request Body/Query
{
"user_id": "your-user-id"
}Important: Never expose your user ID in client-side code. Always use server-side API routes or secure authentication mechanisms.
Upload a PDF document for processing and indexing.
Headers: X-User-ID: your-user-id (required) Content-Type: multipart/form-data Body (Form Data): file: PDF file (required) document_id: string (optional) title: string (optional) file_size: number (optional) file_url: string (optional)
{
"status": "ingested",
"filename": "document.pdf",
"user_id": "user-123",
"document_id": "doc-456"
}curl -X POST https://api.documind.ai/upload \ -H "X-User-ID: user-123" \ -F "file=@document.pdf" \ -F "document_id=doc-456" \ -F "title=My Document"
Ask a question about your documents and get an AI-generated answer.
Headers:
X-User-ID: your-user-id (required)
Content-Type: application/json
Body:
{
"question": "What is the main topic?",
"user_id": "user-123",
"document_id": "doc-456" (optional)
}{
"answer": "The main topic is...",
"sources": ["document.pdf"],
"passages": ["relevant text passage..."]
}curl -X POST https://api.documind.ai/query \
-H "Content-Type: application/json" \
-H "X-User-ID: user-123" \
-d '{
"question": "What is the main topic?",
"document_id": "doc-456"
}'List all documents for a user.
Headers:
X-User-ID: your-user-id (required)
Response:
{
"documents": [
{
"id": 1,
"user_id": "user-123",
"document_id": "doc-456",
"path": "./data/user-123/document.pdf",
"filename": "document.pdf",
"num_chunks": 25,
"title": "My Document",
"file_size": 1024000,
"file_url": "https://..."
}
],
"count": 1
}Get a specific document by ID.
Headers:
X-User-ID: your-user-id (required)
Response:
{
"id": 1,
"user_id": "user-123",
"document_id": "doc-456",
"path": "./data/user-123/document.pdf",
"filename": "document.pdf",
"num_chunks": 25,
"title": "My Document",
"file_size": 1024000,
"file_url": "https://..."
}Delete a document and its associated vectors.
Headers:
X-User-ID: your-user-id (required)
Response:
{
"status": "deleted",
"document_id": "doc-456",
"vectors_deleted": 25
}