Gatsbi

Overview

Gatsbi is a next-generation AI co-scientist that combines research ideation with full academic writing in a single platform. Rather than simply generating text, it begins by analyzing a research topic to detect existing knowledge gaps and proposing 10 or more original research ideas, each scored for originality and paired with relevant references. Users can then expand any idea into a detailed implementation plan, after which Gatsbi's integrated Deep Research Agent collects and synthesizes evidence to produce a fully formatted manuscript of up to 5,000 words (approximately 12 to 16 pages). The platform supports methodological, experimental, case study, and mixed-methods papers, and generates content that meets the publication standards of EI Compendex, SCI, and SSCI venues. Beyond academic writing, Gatsbi drafts comprehensive patent disclosures including background surveys, novelty analyses, embodiments, and formal claims. All output is exportable in MS Word, LaTeX, or Markdown. Operated by Clouchie Limited and Dalian Yunzhui Technology Co., the platform serves over 100,000 researchers across 150 or more countries.

Features

• Research hypothesis generation: Discovers novel research directions by analyzing gaps in existing literature
• AI paper drafting: Generates complete, structured manuscripts with all standard academic sections
• Systematic literature reviews: Automates SLR workflows from search to synthesis with proper methodology
• Meta-analysis generation: Produces quantitative meta-analyses from published study data
• Patent disclosure drafting: Creates patent application documents from research findings and inventions
• Auto-citations and references: Adds in-text citations and builds reference lists from real academic sources
• Multiple AI model backends: Choose from OpenAI, Anthropic, Google, xAI, or hybrid model configurations
• Diagrams and charts: Generates visual elements including tables, charts, and diagrams for manuscripts
• Desktop app with privacy: Local desktop application for private, offline-capable research work
• Web version access: Browser-based interface for working from any device without installation
• Plugin credits system: 2,000 free monthly credits for extended capabilities and integrations
• Multi-language support: Generate and process research content across 11+ languages

Best For

PhD students and early-career researchers accelerating literature reviews and manuscript drafts, Research teams producing multiple papers simultaneously who need to scale their publication output, Inventors and R&D teams drafting patent disclosures from technical research findings, Academic institutions supporting faculty with systematic review and meta-analysis automation

How It Works

Gatsbi follows a three-stage workflow that mirrors how human researchers actually operate. In the first stage (Innovator), the user inputs a research topic and Gatsbi scans existing literature to detect gaps, then generates 10 or more innovative hypotheses with originality scores and supporting references. In the second stage, the user selects a hypothesis and Gatsbi expands it into a detailed implementation blueprint. In the third stage (Writer), the user triggers paper generation: a built-in Deep Research Agent automatically collects and synthesizes evidence from real sources, then assembles a fully formatted draft with equations, figures, in-text citations, and a bibliography. For patent work, the same flow produces a multi-section disclosure document. The platform supports both a web version and a desktop application. Desktop Pro users benefit from local storage with no activity recording. A Humanizer plugin can be applied post-generation to reduce AI detection signals. The entire pipeline from topic input to exportable draft runs within the platform.