A human-in-the-loop research productivity console that turns ideas into trackable literature, Vibe Coding, and Vibe Research workflows—so you can run multiple projects with clarity and speed. Import literature via Undermind or Zotero, log progress and artifacts, and generate automated reports—all locally with no API keys.

Quickstart View on GitHub
60
Demo Projects
16
Research Domains
0
API Keys Required
2
Pipelines (VC + VR)
📋
Fill Template
📥
Import JSON
⚡
Vibe Coding
🔬
Vibe Research
📊
Track & Report
01 · Setup

Quickstart

Deploy locally and start tracking your research workflow in minutes.

install.sh
# Clone & install
$ git clone https://github.com/Mr-Tieguigui/Vibe-Scholar
$ cd VibeScholar
$ pip install -r backend/requirements.txt
$ npm install --prefix frontend
Dependencies installed
run.sh
# Dev mode (hot-reload)
$ uvicorn backend.app.main:app --port 8007
$ npm run dev --prefix frontend
# → http://localhost:5179

# Production build
$ npm run build --prefix frontend
$ uvicorn backend.app.main:app --port 8007
Server running (dev:5179 + api:8007)
workflow.sh
# Full research pipeline
> Export full CSV (Undermind / Zotero)
> Survey Synthesis → Review Doc
> Innovation & Method Design Prompt
> Generate DesignSpec JSON → Import
> Track VC/VR phases → Reports
Pipeline active
02 · Assets

Templates

Standardized formats to bootstrap projects — no external API dependencies.

Project Template

A YAML schema for new projects. Download, fill with any web LLM, upload the JSON.

  • Project Definition & Scope
  • Pipeline & Evaluation
  • Milestones & Deadlines
  • Vibe Coding / Vibe Research Steps
Download YAML

Literature CSV Formats

Example CSV exports. VibeScholar auto-detects formats from column headers.

  • Undermind: Title, Authors, Venue, Relevance, Topic Match
  • Zotero: Title, Author, Abstract Note, DOI, Date Added
Undermind CSV Zotero CSV
03 · Showcase

Demo Projects

60 generic example projects across 16 research domains. All data is synthetic — no private info.

04 · Ingestion

Literature Workflow

Dual-source ingestion: discover with AI, organize locally, track it all.

⬡ Undermind (Cloud)

Discover new, highly relevant papers via AI deep search.

  • Search a topic on app.undermind.ai
  • Review the generated report
  • Download search results as CSV
Relevance Summary Topic Match Score Abstract

📚 Zotero (Local)

Manage a curated, long-term personal reference library.

  • Organize papers in Zotero collections
  • Add notes and PDF annotations
  • Select items → Export as CSV
Item Type Abstract Note Date Added DOI
Export Full CSV→ Survey Synthesis Prompt→ Review Doc→ Innovation & Method Design→ DesignSpec JSON→ Import → Overview / VC / VR / Timeline
05 · Execution

Vibe Coding Stack

Recommended open-source tools and patterns for high-velocity implementation.

Code Agents & IDEs

  • VS CodeExtensible editor for custom tooling.
  • CursorAI-first IDE with codebase indexing.
  • WindsurfAgentic IDE with Cascade.
  • AiderAI pair programming in terminal.
  • OpenDevinOpen-source AI software engineer.
  • Claude CodeAnthropic's agentic coding tool.

Build & Tooling

  • uvFast Python package installer.
  • RuffFast Python linter in Rust.
  • pre-commitMulti-language pre-commit hooks.
  • PydanticData validation via type hints.

Prompt Patterns

Phase Grouping: Organize VC/VR steps into phases (Foundation → Core → Integration → Polish). Prompt Packs: Use short/expanded prompts per step via "Use Prompt" button. Tools: Windsurf (Cascade), Cursor (Copilot++), Claude Code (terminal agent). System: You are an expert engineer. I will provide a task. 1. Outline the files to be modified. 2. Provide a step-by-step implementation plan. 3. Wait for my approval before writing code. Once approved, output complete files without omitting code for brevity.
06 · Intelligence

Vibe Research Stack

Curated tools for literature ingestion, synthesis, and writing.

Literature & PDF Extraction

  • ZoteroOpen-source reference management.
  • Better BibTeXCitation key management for Zotero.
  • GROBIDML extraction from scholarly documents.
  • NougatNeural OCR for academic documents.

Knowledge & Writing

  • ObsidianMarkdown-based knowledge base.
  • BibLaTeXSophisticated bibliography for LaTeX.
  • PandocUniversal document converter.

Prompt Patterns

Task: Literature Triage Input: [Abstract Text] Output JSON: - core_contribution (1 sentence) - relevance_to_my_project (high/med/low) - key_techniques (list) - should_read_full_paper (boolean)
07 · Connect

Contact Us

Questions, feedback, or collaboration ideas? Feel free to reach out.

✉

Email

876506966@qq.com

Best for formal inquiries, collaboration, and project feedback.

💬

WeChat

Tgy876506966

Best for quick communication and direct discussion.

✨

Work With VibeScholar

Interested in research workflow design, tool feedback, or open-source collaboration?

Contact Me
08 · FAQ

FAQ

What is "Vibe Coding"?
A workflow where you direct an AI agent to write code while you focus on architecture, testing, and steering. VibeScholar tracks these loops rigorously.
Do I need an API key?
No. VibeScholar uses no backend AI APIs. Fill templates using any web LLM (ChatGPT, Claude, Gemini) and upload the result.
Where is my data stored?
Locally in plain text files (YAML, JSON, Markdown). No database required — perfect for Git version control.
How does literature import work?
Export a CSV from Undermind or Zotero and upload via the dashboard. Auto-detection normalizes it into a unified index.
Can I self-host on a server?
Yes — standard FastAPI + React. Run on a VPS or keep it local. Designed primarily for desktop use.
What are demo projects?
Synthetic showcase data across 16 research domains. No private information is used or stored.