Knowledge Base Configuration

After creation, you can manage and configure knowledge bases from the Private Knowledge Base list page.

Each knowledge base card in the list displays:

Field	Description
Knowledge Base Name	Display name and parsing mode label
Description	Purpose description
File Count	Number of uploaded documents
Character Count	Total indexed characters in the knowledge base
Associated Apps	Number of AI applications currently referencing this knowledge base

---

Edit Configuration

Click any knowledge base card to enter its edit configuration page.

Available operations:

Operation	Description
File Upload	Upload additional documents to the knowledge base
Recall Testing	Enter test queries to verify retrieval effectiveness; see Recall Testing
Settings	Adjust retrieval mode and visibility (see below)
Authorization Settings	Configure file access authorization; see File Authorization

The file list shows all indexed documents, including filename, character count, upload time, status, and authorized users. You can enable/disable or delete individual files.

---

Settings

Click the Settings button on the edit configuration page to adjust the knowledge base's core configuration.

Visibility

Controls which team members can see this knowledge base:

Option	Description
Only Me	Visible only to the creator
All Team Members	All team members can view and use it
Custom	Manually specify accessible members or departments

Index Mode

Option	Description
High Quality	Calls the system's default embedding interface; higher query accuracy; consumes Tokens
Economy	Uses offline vector engine and keyword indexing; lower accuracy but no Token cost

Index Enhancement

Enable or disable Index Enhancement here. For details, see Create Knowledge Base - Index Enhancement.

Embedding Model

Select the embedding model used to convert text to vectors, e.g., text-embedding-3-large. This affects the semantic undersing capability of vector retrieval.

Retrieval Settings

Adjust the knowledge base's retrieval mode and parameters:

Retrieval Mode	Description
Vector Search	Generates query embeddings and finds the most semantically similar text segments
Full-text Search	Indexes all words in documents; allows querying any word and returns segments containing those words
Hybrid Search	Runs both full-text and vector search simultaneously with a reranking step; requires a Rerank model API

Configurable parameters for vector search mode:

Parameter	Description
Rerank Model	Re-ranks retrieval results to improve relevance of final output
Top K	Maximum number of segments to return; higher values recall more content
Score Threshold	Minimum similarity score; segments below this value are filtered out

After configuration, click Save in the top right to apply.