A research desk for documents you haven't read yet

Read less.
Know more.

Drop in a paper, a report, or a pile of notes. DexCribes.AI compresses it into a clean summary and pulls out exact, highlighted answers to the questions you actually have — powered entirely by open Hugging Face models.

No account. No API bill. Bring your own free Hugging Face token for smoother runs, or try it token-free.

Four tools, one desk

Each one runs on a specific open model chosen for the job, not a single do-everything black box.

Summarize

Executive summaries

Chunks long text and compresses it into clear, bulleted points with DistilBART.

Ask

Extractive Q&A

Ask a question, get the exact sentence the answer came from, with a confidence score — via RoBERTa‑SQuAD2.

Rewrite

Paraphraser

Rephrases a passage in your own words while keeping its meaning intact.

Assist

Research chat

A conversational assistant for working through a document or a question out loud.

Combine

PDF merge

Stitch up to 100 PDFs into one file, entirely in your browser.

Convert

PDF ⇄ image

Turn images into a PDF, or pull a PDF's pages back out as images.

Workbench

Paste text or upload a document, then run it through the tools below.

Ask me about the document above, or anything else you're researching.

PDF utilities

These run fully client-side — nothing leaves your browser.

Merge PDFs

Combine up to 100 PDF files into one, in the order you add them.

    Images → PDF

    Turn one or more images into a single PDF, one page each.

    PDF → images

    Export every page of a PDF as a PNG file.

    Figure generator

    Describe a diagram or illustration for your work — generated by an open text-to-image model. First run can be slow while the model wakes up.

    How it works

    1. You bring the text. Paste it, or upload a .txt, .md or .pdf file — extraction happens locally in your browser.
    2. A request goes to Hugging Face's free Inference API for the model that tool needs — never to a paid service.
    3. The result comes back to your screen. Nothing is stored on a server; refreshing the page clears everything.

    Free inference has rate limits and occasional cold starts. Adding your own free Hugging Face access token (via the “Token” button, top right) raises your limits — it's stored only in your browser, never sent anywhere but Hugging Face.