Getting Started
Parse Sandbox
A shared trial environment for evaluating Pivotly Parse before you go live. Read the guidelines below before uploading your first document.
Quick Start
How it works
Three steps from document upload to extracted insights.
Upload your document
Upload a PDF. Keep files under 500 pages for best performance. Split large documents into focused sections if needed, such as scope, schedules, pricing, and trades.
Extract text
Select the correct page range and run text extraction. Status updates appear at every step so you always know where you are in the process.
Analyze or ask questions
Run AI analysis on the extracted content, or use AI Chat to ask questions directly. No re-run required once text is extracted.
Status updates are shown at every step. If a process takes longer than expected, check the status indicator before re-submitting.
Sandbox Guidelines
Four things to know before you start
These rules exist to help you get the most out of your evaluation time.
Keep documents under 500 pages
- Scope section
- Schedules
- Pricing and quantities
- Trades and specs
This is a shared environment
- Processing speed may vary
- Delays can happen during peak usage
- Do not benchmark performance here
Your data is always private
- No customer sees another's files
- No customer sees another's results
- Only system resources are shared
Sandbox is not production
Capabilities
What you can do in the sandbox
All core Parse features are available for your evaluation.
Upload PDFs and extract structured text
Identify scope items, dates, and deliverables
Extract and analyze financial terms
Flag missing or unclear information
View results linked back to source pages
Ask questions with AI Chat, no re-run required
AI Model
About the AI used in this environment
The sandbox and production environments run on different AI models.
Older model
The sandbox runs on an earlier generation AI model. We use this to keep the shared trial environment cost-effective and accessible. For most evaluation purposes it performs well, but it may produce less precise results on complex or ambiguous documents compared to the current model.
- Sufficient for testing core extraction workflows
- May be less accurate on dense or non-standard documents
- Not representative of production AI performance
Latest model
Private production environments run on the latest available AI model. This means higher accuracy on complex documents, better handling of ambiguous language, and more consistent structured output across your real project files.
- Latest model, updated as improvements are released
- Higher accuracy on complex commercial documents
- Available in your own private Parse environment
Interested in a private environment with the latest model? Reach out to your Pivotly contact to discuss options.