#3838th of 60 in Data and scraping for AIFor data and documents, documents and ocr, science, health and research, research and papers
grobid
Extracts structured TEI/XML from scientific PDFs, including metadata and references
- Stars
- 5.2k
- License
- Apache-2.0
- Last commit
- Nov 2026
- Last release
- Aug 2026
- Language
- Java
4,771 on 22 Apr 20265,227 on 11 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
GROBID parses scientific and technical PDFs into structured XML/TEI using a cascade of sequence-labelling models. CRF models are the default, and DeLFT deep learning models can be enabled via JEP. It extracts headers, references, citation contexts, full-text structure, funders and PDF coordinates, and exposes them through a web service, batch commands and a Java API.
Who it is for: Developers and researchers extracting metadata and text from scholarly PDFs
Strengths
- Reference parsing around .87 F1 on a PubMed Central set with the deep learning model
- Returns PDF bounding-box coordinates for extracted structures
- Reported 10.6 PDF/s full-text on one 16-CPU, 32 GB machine with the Node.js client
- Default CRF models run on commodity CPUs without a GPU
Weaknesses
- Deep learning models are off by default and need Python/JEP, ideally a GPU
- Windows is not currently supported
- Aimed at scholarly PDFs, not general document-to-markdown conversion
- Batch command lines are single-threaded; clients are needed for scale
What it needs
- GPU optional
- Docker
- Needs OpenJDK 21 (source builds), Python 3.10-3.11 with JEP (deep learning models), DeLFT, pdfalto, biblio-glutton or CrossRef API (optional consolidation)
- Models: CRF, RNN, transformers (via DeLFT)
- port 8080
What changed
- 0.9.1 ↗ (opens in a new tab)Security fix
GHSA-MGXF-7MG7-QPMF, GHSA-G2R5-4C8R-C84F, GHSA-47QP-HQVX-6R3F, GHSA-2R2C-CX56-8933, CVE-2026-54513. “Prevent command injection through crafted PDF file names in the non-server pdfalto path: the command is no longer interpolated into a bash -c string but passed as positional parameters and exec'd via…”
- 0.9.0 ↗ (opens in a new tab)Security fix
“Updated Dropwizard to address Trivy vulnerability in Docker image”
- 0.8.2 ↗ (opens in a new tab)Security fix
“Fixed various security vulnerabilities #1125 #1123 #1205”
From the release notes. No flag means the note does not say so, not that the upgrade is safe.
Also in Data and scraping for AI
See all 60| Rank | Project | Score |
|---|---|---|
| 1 | doclingPython library that converts PDFs, Office files and more into structured text | 92 out of 100 |
| 2 | mineruParses PDFs, images and Office files into Markdown and structured JSON | 91 out of 100 |
| 3 | opendataloader-pdfPDF parser that outputs Markdown, JSON with bounding boxes, and Tagged PDF | 91 out of 100 |
| #4 | agent-reachCLI that gives coding agents read access to login-gated and Chinese web platforms | 89 out of 100 |
| #5 | markitdownPython library and CLI that converts office files, PDFs and audio to Markdown | 87 out of 100 |