S SciNet API
Get Token Health Check GitHub

Hosted REST API

SciNet API Documentation

Paper discovery, author retrieval, support-paper mining, and research workflows powered by a scientific knowledge graph.

At a glance

12documented endpoints
JSONagent-friendly I/O
KGgraph-aware retrieval
Tokenpersonal access

Overview

SciNet API is a hosted REST service for scientific knowledge-graph retrieval. It is designed for AI agents, research tools, command-line clients, and scientific workflow systems.

Use it to retrieve papers from structured research plans, discover related authors, look up author publications, mine support papers, and query token status or usage.

Graph-aware retrieval

Follow keywords, citations, paper relations, and author links instead of relying on plain keyword matching.

Agent-ready JSON

Structured request and response shapes make the API easy to call from CLI tools and AI agents.

Personal access control

Email-verified tokens support status checks, usage tracking, and public API access.

Base URL

http://scinet.openkg.cn

Authentication

Most endpoints require a valid SciNet API token. Use the Authorization header when possible; X-API-Key is kept for compatibility.

Authorization: Bearer YOUR_SCINET_TOKEN
X-API-Key: YOUR_SCINET_TOKEN
Public endpoints include health check, browser registration, sending email verification codes, and token registration.

Getting a token

Open the registration page, complete email verification, and copy the returned scinet_xxx token. The token is shown only once, so store it securely.

API Endpoints

Browse endpoint details, copy code examples, or send the selected request from the playground.

Common Request Objects

The high-level search endpoints share a structured research plan and retrieval options.

Plan object

FieldTypeDescription
query_textstringNatural-language topic, idea, or search intent.
source_typestringInput type, such as idea_text, paper_title, or topic.
keywordsarrayWeighted keyword anchors, for example high:open world agent.
titlesarrayKnown paper-title anchors.
reference_titlesarrayReference-paper anchors used to guide retrieval.

Options object

FieldTypeDescription
top_kintegerNumber of main results to return.
retrieval_modestringkeyword, semantic, title, or hybrid.
ranking_profilestringprecision, balanced, discovery, or impact.

Error Handling & Rate Limits

Errors use regular HTTP status codes plus a short error type in the response body.

StatusError typeMeaning
400BadRequestInvalid payload or verification code.
401MissingApiKey / InvalidApiKeyMissing or invalid token.
403TokenDisabled / TokenExpiredDisabled or expired token.
429QuotaExceededDaily quota exhausted.
500ServerErrorBackend or email service error.
Each personal token has a daily quota. Heavier retrieval endpoints may consume more request units than lightweight status endpoints.