Report Insight Lens
Status: Live.
Your muse
Rebuild ofLlama Best Practices: Unlock Rich Insights from Rich Documents (opens in a new tab)Clean
Upload a PDF report and pull out the key numbers, tables and charts. This rebuilds Meta's video "Llama Best Practices: Unlock Rich Insights from Rich Documents" (https://dev.meta.ai/resources/videos/unlock-rich-insights-from-rich-documents) on current Muse models. It demonstrates: Extracting data from PDFs, invoices and reports with the DocumentLens recipe. Plan: Upload a PDF report. Spark 1.3 reads it directly and answers with tables and page references, and SAM 3.1 outlines the charts and figures on page images.
Checks passed
The plan: Report Insight Lens
Upload a PDF report and get key numbers, tables and chart locations.
- Upload a PDF report or invoice up to the size limit
- Ask what numbers or tables to pull or use the default summary
- Review extracted tables with page references and key numbers
- View page images with outlined charts and figures
See the full plan
Who it's for
Analysts, students and small business owners who work with PDF reports and invoices.
How it works
- Upload a PDF report or invoice up to the size limit.
- Ask what numbers or tables to pull or use the default summary.
- Review extracted tables with page references and key numbers.
- View page images with outlined charts and figures.
- Copy or download the results for use elsewhere.
Screens
- Upload Report. Upload a PDF and enter an optional question about numbers or tables.
- Insight Results. Show extracted tables, key numbers and page references.
- Chart Map. Show report page images with outlined charts and figures.
What goes in
- PDF: Business report, invoice or research PDF. Larger files upload to temporary storage first and are deleted after 24 hours.
- Text (optional): Optional question about numbers or tables. Example: total revenue and top expenses.
What comes out
- Text: Tables and key numbers with page references in readable form.
- Overlay: Boxes and outlines over page images marking charts and figures.
- JSON: Structured extract with tables, values and page numbers for copy or download.
Models
Model Used for Why Muse Spark 1.3 Read PDF report directly and extract tables Spark reads PDFs and returns structured tables with page references and key numbers in one call. SAM 3.1 Outline charts and figures on page images SAM finds chart and figure regions from short prompts so pages can show visual outlines. Limits
- Rebuild fit: clean. Rebuilds the same PDF extraction demo on current Muse models without the original footage, script, DocumentLens code, or Meta and Llama branding.
- PDF requests are capped at about 4.5 MB. Larger files use temporary storage and are deleted after 24 hours.
- Page references and numbers are model estimates. Check the original PDF before business use.
- Chart outlines need simple prompts like chart or figure and can miss small or overlapping graphics.
- Built with Muse Muse and is not an official Meta product.
From your inspo
Rebuild of Meta video Llama Best Practices: Unlock Rich Insights from Rich Documents keeping PDF upload and extraction of key numbers, tables and chart locations.
Takes
- Extracting data from PDFs, invoices and reports with tables and page references
- Outlining charts and figures on report pages
Leaves out
- Video footage
- Video script
- Meta and Llama names and logos
- DocumentLens recipe code
Cost per run: about $0.06 (est.). Spark reads about 30k input tokens at $1.25 per million for $0.0375 plus about 1800 output tokens at $4.25 per million for $0.0077. SAM outlines 3 page images at about $0.002 each for $0.006. Total about $0.051, rounded to $0.055 per run. Total $0.055
Recorded demo Done.
npm run verifyexits 0, with all six steps green.