Overview
Sparx is an AI powered document intelligence platform that enables natural language querying
of PDF documents using advanced Retrieval-Augmented Generation technology. Upload your
documents and ask questions to get accurate, sourced answers with page citations.
📄 PDF Processing
Automatic document chunking and vectorization for optimal retrieval
🔍 Smart Search
Semantic search across multiple documents with relevance ranking
💬 Natural Language
Ask questions in plain English and get contextual answers
📚 Source Citations
Every answer includes source document and page number references
🔐 Multi-Tenant
Complete data isolation between tenants with secure authentication
⚡ Streaming
Real-time streaming responses for better user experience
Getting Started
1. Provision Your Instance
Create a Sparx instance through IBM Cloud. Upon provisioning, you'll receive:
- Unique API Key for authentication
- Service endpoint URL
- Dashboard with credentials
2. Upload Documents
Upload PDF documents using the upload endpoint with your API key.
3. Query Your Data
Ask natural language questions and receive AI-generated answers with source citations.
API Endpoints
Base URL https://ibm-sparx.squared.ai
GET
/health
Health check endpoint to verify service status
POST
/upload
Upload PDF document for processing and indexing
GET
/documents
List all documents for your tenant
DELETE
/documents/{id}
Delete a specific document by ID
POST
/chat/completions
Query documents with natural language questions (streaming response)
Usage Example
Upload a Document
curl -X POST https://ibm-sparx.squared.ai/upload \
-H "X-API-Key: YOUR_API_KEY" \
-F "file=@document.pdf"
Query Your Documents
curl -X POST https://ibm-sparx.squared.ai/chat/completions \
-H "X-API-Key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"messages": [
{"role": "user", "content": "What are the key findings?"}
],
"stream": true
}'
Response Format
Streaming Server-Sent Events (SSE) with JSON payloads containing:
- AI-generated answer based on document content
- Source citations with document name and page numbers
- Confidence indicators
- Relevant context from retrieved chunks
Authentication
All API requests require authentication using your tenant-specific API key.
Include the key in the X-API-Key header for all requests.
X-API-Key: YOUR_API_KEY
Multi-Tenancy
Sparx provides complete data isolation between tenants. Each provisioned instance creates
a separate tenant with:
- Dedicated API key linked to the tenant
- Isolated document storage
- Tenant-scoped queries and retrieval
- Independent vectorized embeddings
Your data is never shared with other tenants and queries only access your own documents.
Supported Documents
Sparx works with PDF documents containing text content.
- PDF format (text-based documents)
- Maximum file size: 50MB per document
- Unlimited number of documents per tenant
Service Limits
- Up to 1000 API requests per minute
- Unlimited document storage
- Real-time processing and indexing