Knowledge Base
The Documents area lets you manage the content an agent can retrieve while answering questions.
What you can do
- Add documents from files
- Add content from URLs
- Add manual chunks
- Search existing documents
- Check training and indexing progress
- Inspect chunks
- Review discovered links for URL imports
- Retrain a document
- Configure auto-retrain when supported
- Delete documents you no longer need
Document sources
| Source | What it does |
|---|---|
| Files | Uploads local files for processing |
| URL | Pulls content from one or more web pages |
| Chunks | Adds manually prepared text blocks |
Status labels you may see
| Status | Meaning |
|---|---|
| Progress | The document is being processed |
| Chunking | The content is being split into smaller sections |
| Training | The system is preparing the document for retrieval |
| Indexing | The content is being stored for search |
| Trained | The document is ready to be used by the agent |
| Error | Processing failed and needs review |
tip
Use auto-retrain for sources that change over time, such as documentation pages or knowledge hubs updated on a schedule.