Skip to main content

❓ YingCore FAQ

👤 Best for: All users

⏱️ Reading time: about 6 minutes

💡 In one sentence: The 30 most common questions about running the YingCore platform — read once, skip the pitfalls.

A roundup of the 30 questions that come up most often when administering YingCore, covering eight areas: login, models, knowledge base, workflows, graph, API, security, and digital employees.

Getting started and login​

No account yet — how do I start?​

Ask your admin to open an account → go to http://122.114.233.81/login → log in.

Why can't I see Security Audit / YingClaw / Configuration?​

Those three are admin-only. Ask your admin to adjust your permissions.

Forgot your password?​

Use "Forgot password" on the login page, or ask your admin to reset it.

Where is the platform login?​

http://122.114.233.81/login

Models and AI capabilities​

AI features don't work?​

99% of the time it is because no chat model is configured under Model Providers. Configure at least one chat model first.

How do I connect a self-hosted model?​

Model Providers → pick OpenAI-API-Compatible → fill in base_url (Ollama, VLLM, etc. all work) → save, the model is synced automatically.

Which models do you recommend?​

  • Chat: at least one chat model (DeepSeek, Tongyi, OpenAI all work)
  • Semantic retrieval: an embedding model (one is set up by default)
  • Accuracy boost: a rerank model
  • Image understanding: image2text
  • Voice: speech2text + tts

Knowledge base​

Upload has been "parsing" forever?​

  • Check the progress log (errors are highlighted)
  • Try a different parsing method
  • Use "Re-parse" to rerun (full / incremental)
  • Make sure the file is not corrupted and the format is supported

The AI doesn't cite my documents in its answers?​

  • Did the assistant @ the right knowledge base
  • Is parsing finished
  • Try "Retrieval Test" to confirm recall works
  • Tune the similarity threshold, enable Rerank
  • Tables / images: enable the matching retrieval dimension

How do I exclude a document from retrieval?​

Turn off the "Enabled" toggle next to the document in the list.

Which file formats are supported?​

PDF, Word, Excel, PPT, images, audio, web pages, and more.

Workflows​

A workflow run failed — how do I debug?​

  1. Open "Live Monitor" and find the failing node
  2. Expand it to read the error
  3. Tick two runs in "Run History" and compare
  4. Adjust parameters and rerun

A scheduled workflow never triggered?​

  • Is the schedule switch on
  • Is the cron expression correct
  • Has it hit the max_runs cap

API integration​

API returns 401 Unauthorized?​

Check the Authorization: Bearer <API_KEY> header and confirm the key has not expired.

Error: "model does not exist"?​

The model field must be a real ID from "Get API → Model List", and that model must be configured in Model Providers.

Lost the API Key?​

The Key is only fully visible when created. If you lost it, delete and recreate.

What is the API Base URL?​

https://122.114.233.81/platform-api/v1

Wanxiang graph​

The graph is empty?​

That is normal on first use. Go to Tools → Management Tools → Generate Demo Data to import a sample, or ETL your real data.

Can I delete built-in types (Person, etc.)?​

No, but you can edit their properties. Custom types can be deleted.

Which data sources can I import?​

  • GraphRAG knowledge base
  • SQL connectors (PostgreSQL / MySQL / SQL Server)
  • External HTTP API
  • Direct MySQL connection
  • Bulk import (JSON)

Digital employees and marketing​

How do digital employees and skill packs relate?​

  • Skill pack: a downloadable capability module (from the digital employee marketplace)
  • Digital employee: an AI role that can be orchestrated inside a workflow
  • They work together: install a skill pack → pick the employee in the workflow

AI selection assistant doesn't work?​

It depends on a chat model. Configure one in Model Providers and it will start working.

What is the "Running" plan used for?​

Aggregates stats on the overview page and triggers related digital employee workflows.

How do I embed the AI assistant into my own website?​

Workbench app card → "Embed on site" → copy the iframe code → paste it into your page.

How do I back up data?​

  • Knowledge base: document download / sync tool
  • Graph: "Export JSON" from the canvas / management API
  • Conversations: export from YingClaw

Which languages does the platform support?​

Switch between Simplified Chinese and English from the top-right.

Does the platform support a dark theme?​

Yes. Toggle it from the top-right toolbar.

Security and audit​

Wazuh won't open?​

You see "Failed to fetch Wazuh redirect link": ask your admin to check the Wazuh service and shared key. Admin only.

How do I control what a client can do?​

YingClaw → Request Control → pick a global mode, and add entries to the global sensitive-word filter.

Where do I download the YingClaw client?​

Top toolbar → "Download YingClaw". On macOS, allow it in the security prompt the first time you launch.