research-os 1.3.3

MCP server for reproducible, grounded, citation-verified research — sub-task pipelines, provenance sidecars, publication-grade visualisation, grounded reasoning, and self-tested dashboards.

Research OS

Talk to your AI in plain English. Get publication-grade research back.
No hallucinated citations. No fabricated numbers. Every figure traceable to the script that made it.

Quick start · What you can ask for · Full guide · FAQ

---

What is Research OS?

It's a layer that sits between you, your data, and whatever AI coding assistant you already use (Claude Code, Cursor, Claude Desktop, Antigravity, VS Code, Windsurf, OpenCode, Continue, Aider).

You drop files in a folder. You chat with your AI. Research OS quietly makes sure the AI:

• doesn't make up citations — every paper cited is verified online
• doesn't make up numbers — every figure in your paper traces back to a real script and a real dataset
• doesn't write one giant unreviewable script — work is broken into small, cached, reproducible steps
• doesn't lose track — your decisions, hypotheses, dead-ends, and drafts are recorded so the next chat picks up where this one ended

You get the speed of AI. The work is honest enough to publish.

---

What it can do for you

You don't memorise commands. You just describe what you want.

You say…	You get…
"fill out the intake"	The AI reads your files, proposes a research question + hypotheses, and asks you to confirm.
"what should I do next?"	A short recommended next step — sourced from your state, the literature, and any dead-ends already logged.
"run a baseline EDA"	A numbered experiment folder with scripts, figures (with captions), tables, and a written conclusion.
"compare logistic regression and gradient boosting"	A head-to-head benchmark — same eval, same metrics, paired test, clear winner.
"draft the discussion"	A draft Discussion section that cites your actual results — not invented ones.
"build me a dashboard"	A single-file HTML dashboard. Print-friendly, colour-blind safe, plays offline.
"make a poster for next week"	A LaTeX conference poster with QR code, ≥300 DPI figures, one headline message.
"is this ready to submit?"	A GREEN / YELLOW / RED verdict with a punch list — every check journals run, before they run it.
"going to lunch"	A clean handoff so tomorrow's session resumes exactly where you left off.

→ Full catalogue: USE_CASES.md (by role · by goal · by output type)

---

What you actually see

You open your AI IDE. The MCP server auto-connects. Your folder looks like:

my-project/
├── inputs/             ← you drop files here
│   ├── raw_data/         (CSVs, parquet, FASTQ, …)
│   ├── literature/       (PDFs of papers)
│   └── context/          (notes, Slack threads, PI emails)
│
├── workspace/          ← the AI works here
│   ├── 01_baseline_eda/   (each experiment in its own numbered folder)
│   ├── methods.md         (what was done, in plain English)
│   └── logs/              (every audit + every quality-gate bypass)
│
└── synthesis/          ← your finished deliverables land here
    └── paper.md           (or poster.pdf, dashboard.html, slides.pptx…)

You touch inputs/. The AI touches workspace/ and synthesis/. Things only get created when you ask for them, so a fresh project isn't cluttered with empty folders.

→ Deeper tour: RESEARCHER_GUIDE.md

---

How to use it — 60 seconds

1. Install once (the server is global — it serves every project):

pip install "research-os[all]"

2. Make a project:

mkdir my-project && cd my-project
research-os init                  # answer the arrow-key wizard

The wizard asks where the project goes, which AI IDEs you use (it drops the right MCP config in each), and lets you paste PDFs, DOIs, or notes straight in.

3. Open the folder in your AI IDE and just talk:

> what's in my inputs folder?
> fill out the intake
> run a baseline analysis on the patient data
> what should I do next?

That's the whole workflow.

→ Full first-hour walkthrough: START.md → Per-IDE wiring details: SETUP.md

---

Why use it

Research OS exists because AI coding assistants are fast but they cheat — they invent citations that don't exist, they hallucinate p-values, they write 400-line scripts you can't audit, and they lose track of what they decided last week. The honest fix isn't a better prompt; it's a system that won't let any of those things land in your final paper.

The problem	What Research OS does about it
AI invents citations	Every cite is verified against Crossref / Semantic Scholar / PubMed / arXiv. Unverifiable ones are dropped.
AI invents numbers	Every quantitative claim in your paper traces back to a real workspace file. Synthesis blocks if anything is ungrounded.
AI writes massive unreviewable scripts	Work is split into atomic, content-hash-cached sub-tasks. Edit one; only the affected parts re-run.
AI loses context between sessions	Decisions, hypotheses, dead-ends, and drafts persist. Tomorrow's chat picks up where today's ended.
Figures drift from their captions	Every figure ships with a caption, a plain-English summary, a provenance sidecar (script + seed + library versions), and an SVG companion.
"Pre-registered analysis" silently changes	The Statistical Analysis Plan is content-hashed. Every deviation surfaces at synthesis time.
Negative results vanish into the file drawer	First-class workflow for refuted / underpowered / abandoned findings.
Pre-submission anxiety	Final check walks every gate a journal will run — verdict + actionable punch list.

---

Documentation

If you're…	Read this
New here	docs/START.md — install + first project in 15 min
Looking for an example	docs/USE_CASES.md — what to say for what you want
Going deep	docs/RESEARCHER_GUIDE.md — the full workflow guide
Wiring an IDE	docs/SETUP.md — Claude Code, Cursor, VS Code, etc.
Stuck	docs/FAQ.md — common questions
Curious about a protocol	docs/PROTOCOLS.md — every workflow with triggers + quality bars
Looking up a tool	docs/TOOLS.md — every MCP tool with examples
Sharing a finished project	docs/SHARING.md — share-safe zip + GitHub paths
Contributing a protocol	docs/PROTOCOL_DOCTRINE.md — the scaffold-not-script principle
Driving the AI side	docs/AI_GUIDE.md — what the AI itself reads

Doc index: docs/README.md

---

Tune how the AI behaves

A single file — inputs/researcher_config.yaml — controls everything. Every field is optional. The two that matter most:

interaction:
  quality_gate_policy: enforce        # enforce | allow_override | warn_only
  ambiguity_posture: ask_when_uncertain  # vs take_best_default

→ Full reference: RESEARCHER_GUIDE.md § config

---

Contributing

Issues and PRs welcome.

• Found a bug? → Open an issue
• Want a feature? → Request one
• Code contribution? → CONTRIBUTING.md covers the workflow, branch model, and test conventions.
• Asking a question? → GitHub Discussions
• Security report? → SECURITY.md

---

License

MIT · Research OS does not collect telemetry · Research OS does not manage LLM provider keys (your AI IDE owns model access).

If you use Research OS in published work, citation info lives in CITATION.cff.