Back to the archive

The Reading Room / Deployed

From source document to readable brief

A document reading desk that extracts PDF and image text in the browser, then produces structured summaries and key takeaways at the reader's chosen depth.

Next.jsReactTypeScriptPDF.jsTesseract.jsDeepSeek
* CHECK
SUBJECT
The Abstract
TYPE
Browser-first document AI
STATUS
Deployed

* 3 depths brief, standard, or extended summaries.

PROOF SIGNAL

3 depthsbrief, standard, or extended summaries

The mission

What the system does

A document reading desk that extracts PDF and image text in the browser, then produces structured summaries and key takeaways at the reader's chosen depth.

  • Local PDF parsing and image OCR without uploading the original file
  • Source-text preview alongside structured briefs and key points
  • Map-reduce summarization for documents that exceed a single prompt
  • Copy, regenerate, or download the completed brief as Markdown

Engineering log

Decisions behind the build

What the system handles, and why it works this way.

BOSS 01

Keep the source visible

Extraction and summarization are separate steps. Readers can inspect the recovered text before requesting a brief.

BOSS 02

Long documents

Chunked summaries are synthesized into a final brief, keeping a bounded request flow instead of truncating the source silently.

System map

How signals move

A plain technical view. The map still works without hovering.

01
SourcePDF / image

Local file selection

02
ExtractionPDF.js / OCR

Browser text recovery

03
SummaryNext.js / DeepSeek

Chunking and synthesis

04
BriefReact

Review and Markdown export

LOG ENTRY / WHAT I LEARNED

“A useful summary tool gives readers control over detail and keeps the source close enough to check.”

Source signal

Inspect the work.