Massachusetts Institute of Technology
Principal (Agentic) Investigator
Built by the Poggio Lab team at MIT
Try pAI/MSc, a multi-agent AI system that turns research hypotheses into high-quality, literature-grounded, experiment-backed papers with minimal human steering.
About the Project
Current AI systems require on the order of 102 to 103 prompts to go from a research idea to a written paper. pAI aims to reduce this to up to 10 human interactions.
This is not about fully automating research or replacing human creativity. pAI focuses on the rigorous establishment of ideas: matching theory with experiments, distinguishing correlation from causation, proposing parallel explanations and testing them. The human provides the idea; the system does the structured work of turning it into a solid manuscript.
Getting Started
# 1. Bootstrap environment (one-time, ~5 min) ./scripts/bootstrap.sh researchlab minimal conda activate researchlab # 2. Set your API key (only one provider needed) cp .env.example .env echo "ANTHROPIC_API_KEY=your_key_here" >> .env # 3. Validate setup without spending tokens python launch_multiagent.py --task "test" --dry-run # 4. Run the included quickstart example (~$3) python launch_multiagent.py \ --task "$(cat examples/quickstart/task.txt)" \ --output-format markdown \ --no-counsel \ --no-log-to-files
After the run, look in results/MSc_Student_<timestamp>/ for your generated paper and artifacts.
How It Works
pAI orchestrates 22 specialist agents through a structured research pipeline. Each stage produces artifacts that feed into the next, with human checkpoints for steering.
Search and synthesize relevant papers from ArXiv and the web.
A persona council of 3 critical reviewers debates and refines research directions.
Parallel tracks: prove theorems rigorously and design/run computational experiments.
Synthesize results into a structured, publication-quality manuscript.
Automated proofreading, peer-review scoring, and quality validation gates.
The researchers behind pAI.
Massachusetts Institute of Technology
Corresponding Researcher
Massachusetts Institute of Technology & Perseus Labs
Researcher
Visiting Researcher
Visiting Researcher
Massachusetts Institute of Technology
Senior Researcher
From the Lab
Public Update
Two new ways to run the pAI/MSc research pipeline are now available: a Claude Code skill and an OpenClaw plugin. Both wrap the same core pipeline — 22 agents, 29 backtested prompts, multi-round person...
Public Update
pAI/MSc is now online. You can read the technical report or browse the GitHub repository. This first public release is an open-source, customizable, modular multi-agent system for academic research wo...
