🌐 English · 简体中文
<<<<<<< HEAD 📚 API Docs Website ======= 📚 API Docs Website >>>>>>> 3d1a104 (Update docs and polish downstream frontend outputs)
A pip-installable client and CLI for literature-grounded scientific research workflows on top of the hosted SciNet API.
📄 arXiv · 🔑 Get API Token · 🩺 API Health
SciNet is a research map you can use from the command line. Give it a topic, an idea, an author, or a paper trail, and it helps you look up literature, gather graph-backed evidence, and turn the result into readable reports and reusable JSON artifacts.
Behind that simple workflow is a large scientific knowledge graph. SciNet connects papers, authors, institutions, venues, keywords, citations, and a four-level research taxonomy from domains down to topics. That means a search is not limited to matching words: it can follow how research areas, people, concepts, and papers relate to one another.
This repository packages that capability as a lightweight SciNet client. New users can install it with pip, register an API token, and start running literature-grounded research tasks without setting up Neo4j, maintaining graph data, or touching backend infrastructure.
SciNet spans a broad research landscape, from medicine and social sciences to engineering, computer science, materials science, mathematics, and more.
The graph links papers with authors, institutions, sources, keywords, citations, related work, and the domain-field-subfield-topic hierarchy.
With the client, you can:
- search for papers with keyword, semantic, title, reference, and graph-aware retrieval;
- run research workflows such as literature review, idea grounding, idea evaluation, idea generation, trend analysis, related-author retrieval, and researcher profiling;
- save reproducible outputs such as
request.json,response.json,summary.txt, andreport.md; - customize downstream workflows through editable CLI skills.
- ✨ Overview <<<<<<< HEAD
- 🚀 Quick Start
- 🔑 API Token
- 🧠 What SciNet Does
- 🧩 Supported Tasks
- 🛠️ CLI-First Workflow
- 🧰 Editable Skills
- 🐍 Python SDK
- ⚙️ Configuration
- 🧪 Examples
- 📦 Outputs and Artifacts
- 🛠️ GROBID for PDF Workflows
- 📂 Repository Layout
- 🧯 Troubleshooting
- 🗺️ Roadmap
- ✍️ Citation =======
- � Table of Contents
- 🚀 Quick Start
- 🔑 API Token
- 🧠 What SciNet Does
- 🧩 Supported Tasks
- 🛠️ CLI-First Workflow
- 🧰 Editable Skills
- 🐍 Python SDK
- ⚙️ Configuration
- 🧪 Examples
- 📦 Outputs and Artifacts
- 🛠️ GROBID for PDF Workflows
- 📂 Repository Layout
- 🧯 Troubleshooting
- 📝 TODO
- ✍️ Citation
3d1a104 (Update docs and polish downstream frontend outputs)
Install directly from GitHub:
pip install "git+https://github.com/zjunlp/SciNet.git#subdirectory=scinet"For isolated CLI usage:
pipx install "git+https://github.com/zjunlp/SciNet.git#subdirectory=scinet"After installation:
scinet -hOpen:
http://scinet.openkg.cn/register
Complete email verification and copy your personal token.
Linux / macOS:
export SCINET_API_BASE_URL="http://scinet.openkg.cn"
export SCINET_API_KEY="your-personal-scinet-token"Windows CMD:
set SCINET_API_BASE_URL=http://scinet.openkg.cn
set SCINET_API_KEY=your-personal-scinet-tokenscinet health
scinet configscinet search-papers \
--query "open world agent" \
--keyword "high:open world agent" \
--top-k 3SciNet uses personal API tokens for public access.
Visit:
http://scinet.openkg.cn/register
Steps:
- enter your name, email, organization, and use case;
- click Send code;
- check your inbox for the verification code;
- enter the code and create a token;
- copy the returned
scinet_xxxtoken.
The token is shown only once.
curl -H "Authorization: Bearer $SCINET_API_KEY" \
http://scinet.openkg.cn/v1/auth/token/statuscurl -H "Authorization: Bearer $SCINET_API_KEY" \
"http://scinet.openkg.cn/v1/auth/usage?days=7"SciNet is built for research workflows rather than plain keyword search.
- Search + KG Retrieval: retrieve papers using keywords, semantic matching, title anchors, references, and graph propagation.
- Research Workflow Automation: run literature review, idea grounding, idea evaluation, idea generation, trend analysis, related-author retrieval, and researcher profiling.
- Agent-Friendly Outputs: every run keeps machine-readable JSON artifacts and a human-readable Markdown report.
- Editable Skills: downstream workflows can be represented as JSON skills, inspected, copied, modified, and invoked through CLI.
| Command | Scenario | Main Output |
|---|---|---|
scinet search-papers |
Paper search | Related papers and Markdown report |
scinet related-authors |
Related-author discovery | Candidate authors and scores |
scinet author-papers |
Author paper lookup | Papers by a specified author |
scinet support-papers |
Support-paper retrieval | Evidence papers for candidate authors |
scinet paper-search |
Lightweight low-level paper search | Fast paper candidates |
scinet literature-review |
Literature review | Core paper pool, timeline, writing hints |
scinet idea-grounding |
Idea grounding | Similar works and differentiation evidence |
scinet idea-evaluate |
Idea evaluation | Evidence for novelty, feasibility, and soundness |
scinet idea-generate |
Idea generation | Topic combinations and idea seeds |
scinet trend-report |
Trend analysis | Evolution evidence and representative works |
scinet researcher-review |
Researcher background review | Research trajectory and representative works |
scinet skill |
Editable skill registry | Reusable workflow presets |
SciNet is CLI-first. The CLI is the primary interface for both users and AI agents.
scinet -h
scinet search-papers -h
scinet literature-review -h
scinet skill -hscinet search-papers \
--query "open world agent" \
--domain "artificial intelligence" \
--time-range 2020-2024 \
--keyword "high:open world agent" \
--top-k 3 \
--top-keywords 0 \
--max-titles 0 \
--max-refs 0| Mode | Meaning | Best For |
|---|---|---|
keyword |
Keyword-driven KG retrieval | Clear terminology |
semantic |
Semantic retrieval | Broad semantic matching |
title |
Title-anchor retrieval | Known paper titles |
hybrid |
Keyword + semantic + title + graph walk | Default and recommended |
If --retrieval-mode is omitted, SciNet uses hybrid.
--keyword "high:open world agent"
--title "middle:Voyager: An Open-Ended Embodied Agent with Large Language Models"
--reference "low:JARVIS-1: Open-World Multi-task Agents with Memory-Augmented Multimodal Language Models"| Parameter | Meaning |
|---|---|
--bias-keyword |
Keyword association strength |
--bias-non-seed-keyword |
Non-seed keyword expansion |
--bias-citation |
Citation edge strength |
--bias-related |
Paper relatedness strength |
--bias-authorship |
Author-paper relation strength |
--bias-coauthorship |
Coauthor network strength |
--bias-cooccurrence |
Keyword co-occurrence strength |
--bias-exploration |
Graph exploration level |
--ranking-profile |
Ranking preference: precision, balanced, discovery, impact |
Recommended safe defaults:
--top-k 3
--top-keywords 0
--max-titles 0
--max-refs 0
--bias-exploration lowSciNet skills are JSON presets for downstream research workflows. They make complex workflows easier to inspect, reuse, and customize.
scinet skill list
scinet skill show literature-review
scinet skill run literature-review --query "open world agent" --keyword "high:open world agent"
scinet skill run --dry-run literature-review --query "open world agent" --keyword "high:open world agent"Create a custom skill:
scinet skill init my-review --from literature-reviewThis creates:
./skills/my-review.json
Edit it, then run:
scinet skill run my-review --query "your topic"User-defined skills are loaded from:
./skills/*.json~/.scinet/skills/*.json- directories specified by
SCINET_SKILLS_DIR
User-defined skills can override built-in skills with the same name.
SciNet also provides a lightweight Python client.
from scinet import SciNetClient
client = SciNetClient()
print(client.health())
result = client.search_papers(
query="open world agent",
keywords=[{"text": "open world agent", "score": 10}],
top_k=3,
)
print(result)You can also pass credentials directly:
from scinet import SciNetClient
client = SciNetClient(
base_url="http://scinet.openkg.cn",
api_key="your-personal-scinet-token",
)
print(client.token_status())SCINET_API_BASE_URL=http://scinet.openkg.cn
SCINET_API_KEY=your-personal-scinet-token
SCINET_TIMEOUT=900
SCINET_RUNS_DIR=./runsOptional compatibility variables:
KG2API_BASE_URL=http://scinet.openkg.cn
KG2API_API_KEY=your-personal-scinet-tokenFor new setups, prefer SCINET_*.
scinet literature-review \
--query "retrieval augmented generation" \
--domain "artificial intelligence" \
--time-range 2020-2025 \
--keyword "high:retrieval augmented generation" \
--top-k 5scinet idea-evaluate \
--idea "LLM-based multi-perspective evaluation for scientific research ideas" \
--domain "artificial intelligence" \
--time-range 2020-2025 \
--keyword "high:idea evaluation" \
--keyword "middle:LLM as a judge" \
--top-k 3scinet idea-generate \
--query "knowledge editing for large language models" \
--domain "artificial intelligence" \
--time-range 2020-2025 \
--keyword "high:knowledge editing" \
--keyword "middle:large language models" \
--keyword "low:continual learning" \
--top-k 5scinet trend-report \
--query "retrieval augmented generation" \
--domain "artificial intelligence" \
--time-range 2020-2025 \
--keyword "high:retrieval augmented generation" \
--keyword "middle:knowledge graph" \
--top-k 5scinet researcher-review \
--author "Yoshua Bengio" \
--limit 10 \
--no-abstractTerminal output is concise and table-based. Full outputs are saved under:
runs/<run_id>/
Common artifacts:
| File | Description |
|---|---|
plan.json |
Structured search plan |
request.json |
Full request sent to SciNet API |
response.json |
Raw backend response |
summary.txt |
Short summary |
report.md |
User-facing Markdown report |
metadata.json |
Runtime metadata |
GROBID extracts structured metadata from scientific PDFs, including titles, authors, abstracts, and references. It is only needed for PDF-based workflows.
docker pull lfoppiano/grobid:latest
docker run -d --rm --name grobid -p 8070:8070 lfoppiano/grobid:latest
curl http://127.0.0.1:8070/api/isaliveThen configure:
GROBID_BASE_URL=http://127.0.0.1:8070SciNet/
pyproject.toml
README.md
README_zh.md
README_skills.md
.env.example
src/
scinet/
__init__.py
cli.py
client.py
config.py
skills.py
builtin_skills.json
examples/
search_papers.sh
literature_review.sh
idea_evaluate.sh
tests/
test_import.py
references/
search/
Your token is missing or invalid.
echo $SCINET_API_KEY
export SCINET_API_KEY="your-personal-scinet-token"Windows CMD:
set SCINET_API_KEY=your-personal-scinet-tokenCheck the email address, spam folder, and resend interval.
Use lightweight settings:
--top-k 3
--top-keywords 0
--max-titles 0
--max-refs 0
--bias-exploration lowUse the virtual environment executable directly:
.venv\Scripts\scinet.exe -hor reinstall:
.venv\Scripts\python.exe -m pip install -e .<<<<<<< HEAD
=======
3d1a104 (Update docs and polish downstream frontend outputs)
- CLI Tools. Add more user-facing CLI capabilities so downstream users and AI agents can invoke retrieval workflows without touching database internals.
- Skills. Package reusable agent skills for common scientific discovery workflows and expose best practices as easier-to-load components.
- More Knowledge. Integrate more knowledge forms beyond paper-centric entities, such as datasets, code, standards, theorems, and experimental experience.
- Benchmark and Evaluation. Build dedicated benchmarks and evaluation protocols for downstream scientific research tasks supported by SciNet.
- Dynamic UpdateImprove dynamic knowledge updates toward a more systematic and frequent refresh mechanism.
- Dynamic Update. Improve dynamic knowledge updates toward a more systematic and frequent refresh mechanism.
If you find SciNet helpful, please cite:
This project is licensed under the MIT License. See LICENSE for details.