About BioWrit

โ† Back to dashboard How to use

Overview

BioWrit is an AI writing platform for the biomedical field. From a single input โ€” a paper URL / PMID, a keyword, a NotebookLM notebook, or a manuscript โ€” it produces publication-oriented content in many formats: wiki documents, review articles, original research drafts, editorial comments, AI peer reviews, reviewer response letters, and research proposals.

Under the hood, a Claude-based multi-agent pipeline ingests the sources, generates each output type, and runs quality checks. This dashboard is where you set up a project, send it to the pipeline, and download the results โ€” designed to boost your research-writing productivity.

โš ๏ธ Important notice

This app may produce incorrect results. Every generated draft, citation, and data point must be independently verified by the author, and the author is solely responsible for the final publication. The outputs are reference material to assist writing; they do not replace the author's duty of fact-checking, plagiarism screening, ethical compliance, and final approval.

What is the BioWrit Dashboard

BioWrit is a research-paper writing assistant built on a multi-agent (triangle) system. A Claude master agent (harness) takes an input and orchestrates the pipeline โ€” from source ingestion to document generation and quality checks โ€” while using external models (Codex, ChatGPT) as helpers for code generation and peer-review cross-validation.

This dashboard is the input & status console for that pipeline. The dashboard itself does not generate documents directly; it builds the prompt to hand to harness and tracks whether the outputs have been produced.

How to use

  1. Configure a task โ€” pick output types, enter a keyword (PubMed search) or a NotebookLM URL, then type a new project name or select an existing project.
  2. Build the prompt โ€” click Create task; the harness prompt is generated and saved to the queue.
  3. Run โ€” copy the prompt and paste it into the Claude Code chat to run the harness pipeline. (The dashboard does not launch Claude Code directly.)
  4. Confirm & download โ€” when outputs are produced, the task shows as Done with an alert, and you can download the .docx (wiki as .md).
  5. Revise โ€” select an existing project and issue update instructions to produce a new version.

Output document types

A single project (notebook folder) can hold several output types together.

wiki ยท wiki document (.md) editorial_comment ยท single-paper commentary editorial ยท editor reconciliation review ยท review article research ยท original research article rebuttal ยท reviewer response letter critic ยท AI peer review proposal ยท research proposal
๐Ÿ“Œ Once more

AI-generated results may contain factual errors, inaccurate citations, or outdated information. Review and verify all content before submission or publication. The author bears final responsibility.

About contextBio

contextBio
โ€œ๋ฐ์ดํ„ฐ๋กœ ์ƒ๋ช…์˜ ๋งฅ๋ฝ์„ ์ฝ๋‹คโ€ โ€” Reading the context of life through data

BioWrit is developed and operated by contextBio, an advanced bioinformatics company. contextBio delivers high-dimensional solutions across the research lifecycle โ€” data quality control, primary analysis, interpretation, and integrative analysis:

BioWrit brings this domain-optimized AI to research writing โ€” helping scientists produce rigorous, publication-ready content faster, while keeping the author in control of accuracy and final judgment.

E-mail
contact@contextbio.co.kr
HQ
111 Hoegi-ro, Dongdaemun-gu, Seoul (Hongneung Bio-Medical Startup Center), Room 206-1
Lab
56 Geumto-ro 80beon-gil, Sujeong-gu, Seongnam, Gyeonggi (Widen Tower), Room 211

Copyright ยฉ contextBio ยท Research use only