AI Research OS
A Self-Evolving Research Operating System for AI Researchers
What It Does
AI Research OS is a self-evolving research system that learns from your usage patterns. It's not just a paper manager — it's a research partner that grows smarter over time.
Feed it a paper (arXiv URL, DOI, or PDF). Get back a P-Note, C-Note, Radar entry, and Timeline entry — all structured, tagged, and cross-linked.
| Input | Output |
|---|---|
| arXiv URL/ID | P-Note + C-Note + Radar + Timeline |
| DOI | P-Note + C-Note + Radar + Timeline |
| Local PDF | P-Note + C-Note + Radar + Timeline |
| Scanned PDF | Same (via OCR) |
This is not a PDF manager. It is a Self-Evolving System that:
- Learns from your research patterns
- Improves answers over time
- Adapts to your specific domain
Core Features
| Feature | Description |
|---|---|
airos import |
Import papers from arXiv, DOI, PDF |
airos chat |
RAG-powered Q&A with your papers |
airos slides |
Auto-generate presentations |
airos kg |
Knowledge graph visualization |
| Evolution | Self-improvement via Gene/Capsule patterns |
Quick Start
pip install ai-research-os
airos-cli 2601.00155 --tags LLM,Agent
That's it — one paper imported in seconds. The above installs the package and imports an arXiv paper.
One line, three inputs
airos-cli 2601.00155 # arXiv ID
airos-cli 10.48550/arXiv.2601.00155 # DOI
airos-cli --pdf paper.pdf --tags RAG # Local PDF
airos-cli --pdf scanned.pdf --ocr --ocr-lang chi_sim+eng # Scanned PDF
Three core commands
airos-cli import 2601.00155 10.1038/nature12373 # Add papers to DB
airos-cli search "attention mechanism" --tag LLM # Search papers
airos-cli research "RLHF alignment" --limit 5 # Autonomous research loop
AI draft (optional)
export OPENAI_API_KEY="***"
export OPENAI_BASE_URL="https://dashscope.aliyuncs.com/compatible-mode/v1"
airos-cli 2601.00155 --tags LLM --ai
For full configuration, see API_CONFIG.md.
Research Tree
Papers are organized into 12 directories:
00-Radar/ Topic heat tracking
01-Foundations/ Foundational papers
02-Models/ Model papers
03-Training/ Training methods
04-Scaling/ Scaling laws
05-Alignment/ Alignment research
06-Agents/ Agent systems
07-Infrastructure/ Infrastructure
08-Optimization/ Optimization techniques
09-Evaluation/ Evaluation methods
10-Applications/ Applied research
11-Future-Directions/
Installation
pip install ai-research-os
Or install from source:
git clone https://github.com/shushuzn/ai_research_os.git
cd ai_research_os
pip install -e .
Documentation
Full documentation at ai-research-os.readthedocs.io.
| Doc | Description |
|---|---|
| Architecture | System design and module overview |
| Configuration | LLM, DB, Search, Tool configuration |
| Benchmarks | Performance metrics and test coverage |
| Contributing | How to contribute to this project |
| Roadmap | Project roadmap and future plans |
License
GPL-3.0-or-later. See LICENSE for details.
Metadata
Release files for ai-research-os 1.5.4
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| ai_research_os-1.5.4.tar.gz | 809.8 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| ai_research_os-1.5.4-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 1.4 MB
Release files / ai_research_os-1.5.4.tar.gz
| Download URL | ai_research_os-1.5.4.tar.gz |
|---|---|
| Size | 809.8 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
aa3ef8b1e9f596f81aa367750353bfc290e9f3e3a27896c29868cbb9da5817fc
|
|
BLAKE2b-256 checksum How to use checksums |
edb9d3da9c4505c688e8ef7d530048b546ca349a2e71bed93b0e4b1a3ce7fb12
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.11.15
|
Release files / ai_research_os-1.5.4-py3-none-any.whl
| Download URL | ai_research_os-1.5.4-py3-none-any.whl |
|---|---|
| Size | 584.4 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
a05b93dec31224f25438605275f8ee74c48d7c3baf7f6a3e88a23155321380f2
|
|
BLAKE2b-256 checksum How to use checksums |
e763c3b28c3702b4a27b638c769aa6cca95907e4e2ec8aaa7e5bea4126fcb5d2
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.11.15
|