Transform research papers, book highlights, and technical PDFs into connected Obsidian notes with YAML metadata, KaTeX formulas, and callout boxes.
100% In-Browser Conversion — Your Files Are Never Uploaded
Drag & Drop your PDF document here
or browse files on your device
See how our parser transforms flat PDF text into a fully structured Obsidian note complete with Properties frontmatter, math blocks, and callouts:
Title: Neural Scaling Laws for Autoregressive Language Models
Authors: Jared Kaplan, Sam McCandlish, Tom Henighan, et al.
Date: 2020
Abstract: We investigate empirical scaling laws for language model performance...---
title: "Neural Scaling Laws for Autoregressive Language Models"
authors: ["Jared Kaplan", "Sam McCandlish", "Tom Henighan"]
date: 2020-01-23
tags: [research, ai, llm, scaling-laws]
aliases: ["Kaplan et al. Scaling Laws"]
---
# Neural Scaling Laws for Autoregressive Language Models
## Abstract
We investigate empirical scaling laws for language model performance. Key findings link to [[Transformer Architecture]] and [[Compute Optimal Training]].
### Power-Law Performance Equation
$$
L(N) = \left(\frac{N_c}{N}\right)^{\alpha_N}
$$
> [!NOTE] Key Takeaway
> Cross-entropy loss scales as a power-law with model size $N$, dataset size $D$, and compute $C$.Obsidian's power lies in bi-directional linking, property-based querying (via Dataview), and local plaintext Markdown storage. When storing PDFs purely as binary attachments, their content remains locked in static silos that cannot be queried or interconnected.
Everything you need to know about format extraction, privacy, and markdown compatibility.
Simply drag the downloaded `.md` file directly into your Obsidian vault folder in your OS file explorer (Finder on macOS or File Explorer on Windows). Obsidian instantly indexes the note, populates the YAML frontmatter in properties view, and integrates the note into your interactive Graph View.
Yes. Our engine automatically extracts document title, author metadata, creation dates, and subject tags, formatting them into standard YAML frontmatter (`---` blocks) that populates Obsidian 1.4+ Properties natively.
Yes. All extracted equations are wrapped in standard `$...$` (inline) and `$$...$$` (display blocks), which are rendered natively by Obsidian's built-in KaTeX engine without requiring any third-party community plugins.
Yes. If your PDF contains embedded highlight annotations and sticky notes, our parser extracts them into formatted Obsidian Callout blocks (`> [!NOTE]` or `> [!QUOTE]`) with direct page reference anchors.
Explore dedicated documentation for other document formats and developer pipelines.
Import PDFs with bullet hierarchies and block references into Logseq knowledge graphs.
Extract complex scientific equations and matrices into Obsidian-ready KaTeX syntax.
Prepare Obsidian vault notes for local vector search and Smart Connections AI plugins.