Your coding agent already reads code, writes code, and runs experiments. ScholarAIO adds a structured research workspace on top, so the same agent can search literature, cross-check results against papers, use scientific software more accurately, and carry the whole research workflow from one terminal.
Your paper library becomes a reusable knowledge base for the same agent.
When scientific software questions come up, the agent can consult official documentation at runtime instead of guessing from prompts.
The system is built to keep expanding as new tools and workflows become worth supporting.
ScholarAIO offers more than search. It gives an AI coding agent a research workspace that supports natural-language interaction, papers and notes, more reliable use of scientific software, writing and running code, checking results against the literature, and structured academic writing.
Quick Start
The default and recommended way to use ScholarAIO is simple: install it, configure it once, and open this repository directly with your coding agent.
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
First-time install walkthrough for Claude Code, Codex CLI, and ChatGPT.
Then open the repository in Codex, Claude Code, or another supported agent. In this setup, the agent gets the fullest experience: bundled instructions, local skills, the CLI, and the complete codebase context are all available directly. For Claude Code plugins, Codex/OpenClaw skill registration, and other setup paths, see docs/getting-started/agent-setup.md.
Upgrading To 1.4
ScholarAIO 1.4 is a runtime-layout upgrade. It does not migrate user data
automatically during git pull, pip install -U, or normal CLI startup. That is
intentional: data movement is an explicit offline operation with a migration
journal and verification.
Recommended path:
# 1. Update the code/package
git pull
pip install -e ".[full]"
# 2. From the ScholarAIO runtime root, inspect and migrate explicitly
scholaraio migrate status
scholaraio migrate upgrade --migration-id upgrade-1.4.0 --confirm
scholaraio migrate verify --migration-id upgrade-1.4.0
# 3. Rebuild indexes after migrated data lands in the fresh layout
scholaraio index --rebuild
For the lowest-risk upgrade, keep or copy your old ScholarAIO folder first, then
run the migration in the upgraded checkout that contains your data/,
workspace/, and config*.yaml. See
docs/getting-started/upgrading-to-1.4.md.
What It Does
| | Feature | Details |
|--|---------|---------|
| PDF Parsing | Deep structure extraction | Convert PDFs into structured Markdown while preserving formulas, figures, and layout as much as possible |
| Not Just Papers | More than papers | Journal articles, theses, patents, technical reports, standards, and lecture notes — four inbox categories with tailored metadata handling |
| Hybrid Search | Keyword + semantic fusion | Combine full-text and vector retrieval for stronger search results |
| Topic Discovery | See what your library is about | Automatically group papers into research themes and use interactive views to grasp the overall structure quickly |
| Literature Exploration | Multi-dimensional discovery | Explore a research direction through journal, topic, author, institution, keyword, year, citation impact, and more |
| Citation Graph | References & impact | Forward citations, backward citations, and shared-reference analysis |
| Layered Reading | Read on demand | Start with metadata or the abstract, then move into conclusions or full text only when you need to |
| Multi-Source Import | Connect your existing library | Import directly from reference managers, PDFs, and Markdown without rebuilding your library from scratch |
| Workspaces | Organize by project | Manage paper subsets with scoped search and BibTeX export |
| Multi-Format Export | BibTeX, RIS, Markdown, DOCX | Export your full library or a workspace for Zotero, Endnote, submission, or sharing |
| Metadata Scrub | Incremental cleanup after enrich | Review and repair low-quality titles, authors, and years for non-standard documents, then mark reviewed records to skip future passes |
| Persistent Notes | Cross-session memory | Keep analysis notes for each paper so future sessions can reuse them instead of starting over |
| Research Insights | Reading behavior analytics | Search hot keywords, most-read papers, reading trends, and semantic neighbor recommendations for papers you haven't read yet |
| Federated Discovery | Cross-library search | Search your main library, exploration libraries, and arXiv from one entry point instead of hopping across tools |
| Remote Backup | Rsync-based sync | Back up the ScholarAIO data/ workspace to configured remote targets through named rsync plans |
| AI-for-Science Runtime | Use scientific software more accurately | Use scientific software against official documentation at runtime instead of guessing commands and parameters |
| Extensible Tool Onboarding | Keep adding the tools that matter | As new scientific tools and workflows become important, the system can keep expanding |
| Academic Writing | AI-assisted writing | Router-first workflows for literature review, guided single-paper reading, paper sections, citation check, rebuttal, gap analysis, poster packages, and technical reports — with every citation traceable to your own library |
For writing tasks, start with the router-style writing entry when the deliverable is clear but the workflow is not. The current writing stack is organized around:
academic-writing: route by deliverable and writing stage
literature-review: long-form review and survey writing
paper-guided-reading: guided deep reading of a single paper from fuzzy search to full-text analysis
paper-writing: manuscript sections and paper-focused drafting
review-response: rebuttal and response-letter workflows
research-gap: gap analysis and open-question reports
technical-report: technical briefings and topic reports
ScholarAIO is designed to be agent-agnostic, but different agents expose different integration paths. Some work best when you open this repository directly; others are easier to use through plugins.
Skills follow the open AgentSkills.io standard, and .agents/skills/ and .qwen/skills/ are symlinks to .claude/skills/ so different agents can discover and reuse the same skills. Qwen-specific project context lives in .qwen/QWEN.md.
Migrating from existing tools? Import directly from Endnote (XML/RIS) and Zotero (Web API or local SQLite), with PDFs, metadata, and references brought over together. More import sources are on the roadmap.
Configuration
Start by opening scholaraio with your agent and let it walk you through the setup. The notes below are only a basic overview.
ScholarAIO works with a minimal setup and can be expanded as needed.
scholaraio setup walks you through the basics.
An LLM API key is optional but recommended for more robust metadata extraction and content completion.
A MinerU token is optional but recommended, and free. You can also deploy MinerU or Docling locally for PDF parsing.
scholaraio setup check shows what is installed, what is optional, and what is missing.
ScholarAIO works best through an AI coding agent, but it also provides a CLI for scripting, debugging, and quick queries. For a current command reference aligned with the code, see [docs/guide/cli-reference.md](docs/guide/cli-reference.m
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.
