Core Features
Search and Retrieval
How AINexLayer finds relevant content for chat answers
AINexLayer retrieval is chat-first. There is no separate global knowledge-base search screen in the app. You find information by asking in chat (and optionally narrowing with @ mentions).
How retrieval works
- You ask a question (optionally
@a file, folder, or connector). - The agent retrieves relevant passages from indexed Library documents (and may use tools or web/native scrapers when appropriate).
- The model answers using that context.
- Citations show which sources supported the reply.
This is semantic retrieval: meaning matters more than exact keyword match. Ask in natural language.
Ways to focus results
- @ a document or folder — strongest signal for “use this source”
- Be specific — names, dates, and section titles improve ranking
- Follow up — refine instead of restarting when the first pass is close
- Connect tools — use connectors for live systems that are not in the Library
Searching chats vs knowledge
- Search chats… in the all-chats UI finds past conversations by title/content signals.
- Library filters (for example filter by type) help you browse Documents; they are not a replacement for chat Q&A.
Tip
If you expected a keyword search box over all documents, use chat instead: ask “Where do we mention vacation policy?” and open the citations.