PDF and Knowledge: Multi-Page Report Summary Template

Drop a long PDF report into Spojit and get back a tight executive summary with the key figures pulled out, all from a single Miraxa prompt.

What It Builds

This template starts from a Manual trigger so you can run it on demand with any report PDF. The pdf connector extracts the full text, a Knowledge node embeds that text into a Transient collection scoped to the single run, and a second Knowledge node in Query mode reads back an executive summary with the headline numbers called out. Because the collection is Transient, the document is embedded, queried, and discarded inside one run.

The Prompt

Paste this into Miraxa and it builds the workflow, connecting the tools for you:

Build a manual workflow where I upload a long multi-page PDF report. Extract the text from the PDF, embed it into a transient knowledge collection for this run only, then query it to return a one-paragraph executive summary followed by a bullet list of the key figures (totals, percentages, and dates) found in the report.

Connectors Used

Manual trigger - run it yourself whenever a report comes in.
pdf - extracts text from the multi-page document, including OCR for scanned pages.
Knowledge - one node embeds the text into a Transient collection, a second node queries it for the summary.

Customize It

Reword the prompt to change the output shape: ask for a fixed number of bullet points, a specific tone for the summary, or a section-by-section breakdown. You can also name the exact figures you care about, such as revenue, headcount, or quarter-over-quarter growth, so Spojit surfaces them every time.

Tips

A Transient collection keeps each report isolated; switch to a persistent collection if you want to search past reports later.
Add a Response Schema to the Query node when you need the figures as structured JSON for a downstream node.
Very long reports embed faster when scanned pages are clean; the pdf connector handles OCR but clearer scans give sharper figures.