Knowledge is content you give an assistant to reference when answering questions and completing tasks. Instead of relying solely on general training, your assistant can draw on your specific documentation, policies, guides, and any other context you provide.Documentation Index
Fetch the complete documentation index at: https://docs.quiva.ai/llms.txt
Use this file to discover all available pages before exploring further.
How Knowledge Works
When you add knowledge to an assistant, that content is made available during every conversation. The assistant searches and references relevant knowledge automatically — you don’t need to prompt it to look things up. Knowledge is stored at three levels:Account-Level
Global knowledge shared across your entire account. Enabled per assistant as needed.
Team
Knowledge available to all team members using this assistant.
Personal
Your private knowledge, layered on top of team knowledge. Visible only to you.
Adding Knowledge
Knowledge can be added from three sources:- File Upload
- URL
- Manual Entry
Upload documents directly from your computer.Supported formats:
- PDF documents
- Word documents (.docx)
- Plain text files (.txt)
- Markdown files (.md)
- Open your assistant configuration
- Click the Knowledge tab
- Click Add Knowledge → Upload File
- Select your file and confirm
Account-Level Global Knowledge
Global knowledge is configured at the account level and can be enabled on any assistant. It’s designed for company-wide content that many assistants might need — brand guidelines, company policies, product overviews, or shared reference material. Setting up global knowledge:- Go to Account Settings → Global Knowledge
- Add knowledge sources (files, URLs, or manual entries)
- In each assistant’s Knowledge tab, enable the global knowledge sources you want that assistant to use
Global knowledge isn’t enabled by default on every assistant — you choose which sources each assistant can access. This keeps assistants focused and prevents unnecessary context.
Team vs. Personal Knowledge
Every assistant has two knowledge layers: Team Knowledge is configured by assistant editors and shared with everyone who uses the assistant. It provides the shared foundation — the product documentation, policies, and context that everyone on the team should have access to. Personal Knowledge is your private layer. It’s added and visible only to you, and it layers on top of team knowledge at runtime. Use it to add your own reference material, notes, or context without affecting the experience for other team members.Large Document Processing
QuivaWorks handles documents that exceed the normal context window through automatic sub-agent processing. When a large document is added as knowledge:Document Indexing
The document is broken down into structured sections and stored in memory. This happens once when the knowledge is added.
Retrieval at Runtime
During a conversation, a sub-agent searches the indexed document and retrieves the most relevant sections.
Best Practices
Keep knowledge focused and relevant
Keep knowledge focused and relevant
Only add knowledge that’s genuinely useful for the assistant’s role. Irrelevant knowledge doesn’t help and can dilute retrieval quality. A customer service assistant doesn’t need the engineering team’s architecture decisions.
Use clear, well-structured documents
Use clear, well-structured documents
Documents with clear headings, sections, and structure are indexed more effectively. Avoid dense walls of text — use headers, bullet points, and clear paragraph breaks where possible.
Keep knowledge up to date
Keep knowledge up to date
Knowledge isn’t automatically refreshed when source documents change. When you update a policy or documentation page, re-add the knowledge source to keep the assistant current.
Use manual entries for frequently referenced policies
Use manual entries for frequently referenced policies
For short, important content like return policies or escalation procedures, manual entries are the easiest to manage. They can be edited directly and don’t require re-uploading files.
Use global knowledge for shared foundations
Use global knowledge for shared foundations
Next Steps
Information Settings
Configure instructions, name, and assistant type
Tools & Connectors
Connect to external systems via MCP
Capabilities
Image analysis, file generation, and app deployment
Best Practices
Optimise your assistant’s performance