YingCore Core Features
📖 Best for: Product managers / business owners / first-time decision makers — people who want to know what problems YingCore can solve
📖 Reading time: 3 minutes
📖 In one sentence: YingCore is the server-side "brain" of YingClaw, unifying identity & permissions, skill marketplace, security & compliance, audit trails, and cross-agent orchestration — so every conversation, every file operation, and every scheduled task in the YingClaw client runs in an orderly way.
| Module | One-line Definition | Detailed Capabilities | Typical Use Cases |
|---|---|---|---|
| Marketing Plans | Smart marketing-plan hub: 4 templates x 8 sub-modules for structured generation (brand/product dual-track) | One-stop management + one-click AI full plan generation + selection assistant (rewrite/summarize/translate/unify terminology) | Draft brand plans, product innovation plans, and event plans - let AI be your plan architect |
| Workflow (one of the platform's biggest efficiency boosters) | Set up recurring processes on a schedule - log in daily and the work is already done | Visual canvas (drag-and-drop to assemble flows): chain together "AI call, HTTP request, conditional branch, loop" steps • 4 trigger modes: Manual / Scheduled cron / One-shot timer / Webhook • Strong debugging: test run, breakpoint debug, run-diff comparison | Auto-aggregate yesterday's customer inquiries each morning -> email to sales • Auto-generate last week's marketing data report every Monday • Form submitted by user -> auto-analyze intent -> route to the right rep |
| Knowledge Base (the foundation of AI "understanding your company's business") | A system for organizing industry knowledge | Upload company materials: product manuals, SOPs, training docs, customer contracts, industry reports, etc. • System auto-slices large documents into small chunks • 15 parsing methods: general, Q&A, resume, table, paper, PPT, contract, image, audio, etc. • Multi-dimensional retrieval: in addition to text, tables and images are also searchable | Customer service system: customer asks a question, AI auto-finds the answer in the product manual • HR system: employee asks about leave / reimbursement policy, AI answers directly • Sales system: rep asks for product specs or pricing strategy, AI answers in real time |
| Workspace (AI Q&A) | Multi-turn AI conversation - ask anything | Default assistant is ready out of the box; with the knowledge base, AI answers based on company materials • Answers include "File Citations" - click to see the original source location • Web search, upload files to participate in conversation, expand thinking process, customize assistant persona • 15+ pre-built assistant apps: Xiaohongshu monitor, sales leads, complaint analysis, voice cloning, lip-sync, video subtitles, etc. | Ask AI to polish your weekly report • Analyze trending notes on Xiaohongshu • Auto-generate subtitles for videos • Paste a blogger's profile page and let AI analyze it |
| Model Plaza | Unified model view, smart routing, and 5 MCP services that expose platform capabilities to external clients | • Model grid + smart routing: 12 cards per page showing vendor / status / price / capabilities / latency, 4 stats cards + keyword / type / status filters; 8 task types (Code / Reasoning / Chat / Creative / Summary / Translation / Agent / Default) with primary model + policy, plus auto (sticky-stable) and smart-auto (dynamic intent) virtual routing entries • MCP services: 5 independent instances (Knowledge Base /kb, Data Hub /datahub, Document Image /image, Knowledge Graph /graph, LLM-Wiki /wiki) sharing one Key, keys can bind to accounts for per-caller audit • See the Model Plaza docs for the full guide | Spot red "Blocked" and orange "Cooling" first, route by task to cut cost, wire MCP JSON into Claude Desktop, bind account so audit traces the caller |
| Cloud Workspaces | Multi-user collaborative cloud file and runtime space | • Unified folder structure with fine-grained permission control • Deep integration with Workflow / Knowledge Base / Model Providers • Files can be used directly as AI input for conversations or processing | Team asset library: designer puts assets in cloud space, AI calls them on demand • Multi-project parallel: one workspace per project, files don't interfere • Audit trail: who changed which file when — at a glance |
| Market Center | Task breakdown and tracking hub for marketing plans: 7 modules (Plan Workbench / Plan / Task Tree / Department Dispatch / Metric Dictionary / Org Version / Dispatch & Notification) | Plan templates + 3-level task breakdown (plan -> department KPI -> individual) + department dispatch + unified metric dictionary + org version snapshots + dispatch & notification + exception closed-loop | 618 promotion breakdown, monthly sales KPI dispatch, cross-department product launch, offline action backfill, task migration after org restructure |
| Knowledge Graph (weave people, events, and things scattered across the company into a network - see the relationships clearly) | Company info ingestion and analysis (entity + relationship knowledge graph) | • Ingest: customers, suppliers, products, contracts, projects, employees... define an object type • Relations: Company A "supplies" Product B, Employee C "owns" Project D • Graph analysis: centrality, community detection, path finding, duplicate-entity detection and merging • Multi-source import (Knowledge Base ETL / SQL connector / HTTP API) | Enter all customers, suppliers, contracts into the graph - see the full chain of "who->to whom->supplies what" • Find all related parties of a given customer • Risk monitoring: spot a supplier's affiliate in trouble and get an early warning |
| Data Platform (let business users query data themselves - with sources attached) | Data assets + natural-language queries + federated direct query | • 5 source types (SQL/API/File/NoSQL/Document) via federated direct query, raw data never lands by default • Natural-language queries with source tags • Field authorization matrix + per-query audit (plaintext / masked / denied) • Unified metric / tag / audience / profile assets | Ops asks "last month's East China repurchase rate" — AI answers directly with sources • Same "active user" defined three different ways across three reports — define as metric for one canonical answer • Field-level auth + traceable access for sensitive columns |
| Digital Employees (like hiring a batch of 7x24-hour tireless AI workers) | Handle simple, repetitive work for you | • AI Creator: auto-generate marketing copy, product descriptions, social media content • AI Publisher: auto-adapt and post content to Xiaohongshu, Douyin, and other platforms • AI Analyst: data collection, user behavior analysis, public-opinion monitoring • AI Customer Service: auto-answer customer questions, identify and escalate to humans • Skill Marketplace: download skill packs from the community, Tencent, ModelScope | • Every day at 8 AM AI Creator auto-writes the day's marketing copy; at 9 AM AI Publisher auto-posts to Xiaohongshu • Customer leaves a sales question in a Douyin comment -> AI Analyst auto-identifies the lead and routes it to a sales rep • AI Customer Service is on duty 7x24, answering simple questions directly and escalating complex ones to humans |
| Security Audit | Entry point to Wazuh for security monitoring | • Top nav menu, super admin only (JWT SSO, no password needed) • 5 modules: HIDS / log analysis / file integrity monitoring (FIM) / vulnerability detection / alerts and response • Covers host intrusion, file changes, vulnerabilities, cross-host log tracing • Strictly separated from YingClaw Session Audit (each handles its own scope) | Host CPU spikes -> HIDS alert -> drill into logs and file changes • Critical file (e.g. /etc/passwd) change alert • Bulk vulnerability remediation: scan 50 hosts for the same CVE -> schedule fix -> re-scan to confirm |
Next Steps
- YingClaw Core Features - see what the client-side digital employees can do
- How to Get Started - first-time login and how to give it a task
- Common Questions - the questions people ask most often