Multi-Modal Research Agent

Upload images and PDFs to analyze charts, extract data, and discover insights across multiple documents

📊 Compare Charts

Upload quarterly report charts and compare trends across periods

📝 Extract Tables

Pull data from image screenshots and cross-reference with PDF reports

🔍 Verify Data

Check if data in presentation slides matches source documents

Multi-Modal Research Agent

Drop files anywhere or use the + button to upload images and PDFs

Technical Highlights

Multi-Modal Message Parts

AI SDK v6's parts array allows mixing text, images, and PDFs in a single message. Files are converted to data URLs and sent as file parts alongside text content.

GPT-5.1 Thinking Vision Capabilities

Uses gpt-5.1-thinking model with native vision support to analyze images and PDFs. Model can extract text, interpret charts, and identify visual patterns across document types.

Cross-Document Analysis

System prompt guides the model to compare data between files, identify relationships, and highlight discrepancies or confirming patterns with source citations.

Try These Examples:

1.Upload a chart image + quarterly report PDF → Ask: "Do the numbers in this chart match the Q4 data in the report?"

2.Upload multiple product screenshots → Ask: "Compare the UI designs and identify common patterns"

3.Upload invoice PDF + receipt photo → Ask: "Verify these are for the same transaction and check if amounts match"

Featuring Multi-Modal Messages + GPT-5.1 Thinking Vision

📸 Images📄 PDFs🔗 Cross-Reference

Multi-Modal Research Agent

Upload images and PDFs to analyze charts, extract data, and discover insights across multiple documents

📊 Compare Charts

Upload quarterly report charts and compare trends across periods

📝 Extract Tables

Pull data from image screenshots and cross-reference with PDF reports

🔍 Verify Data

Check if data in presentation slides matches source documents

Multi-Modal Research Agent

Drop files anywhere or use the + button to upload images and PDFs

Technical Highlights

Multi-Modal Message Parts

AI SDK v6's parts array allows mixing text, images, and PDFs in a single message. Files are converted to data URLs and sent as file parts alongside text content.

GPT-5.1 Thinking Vision Capabilities

Uses gpt-5.1-thinking model with native vision support to analyze images and PDFs. Model can extract text, interpret charts, and identify visual patterns across document types.

Cross-Document Analysis

System prompt guides the model to compare data between files, identify relationships, and highlight discrepancies or confirming patterns with source citations.

Try These Examples:

1.Upload a chart image + quarterly report PDF → Ask: "Do the numbers in this chart match the Q4 data in the report?"

2.Upload multiple product screenshots → Ask: "Compare the UI designs and identify common patterns"

3.Upload invoice PDF + receipt photo → Ask: "Verify these are for the same transaction and check if amounts match"