Knowledge Base (RAG)

Per-app pgvector knowledge bases: ingest documents, search, and chat with grounded citations.

Endpoints

Every app gets an isolated vector knowledge base (pgvector). Ingest text or URLs, then search or chat with grounded citations — billed through the same API key. The KB service lives on its own host: https://router-kb.intelli-verse-x.ai.

POST/v1/kb/ingest

POST/v1/kb/search

POST/v1/kb/chat

Ingest documents

Parameters
Name	Type	Description
`app_id`Required	string	The app whose knowledge base receives the documents.
`documents`Required	array	Items with either `url` (fetched and chunked) or `text` + optional `title`.

Shell

curl https://router-kb.intelli-verse-x.ai/v1/kb/ingest \
  -H "Authorization: Bearer $INTELLIVERSE_API_KEY" \
  -d '{ "app_id": "APP_ID", "documents": [{ "url": "https://docs.myapp.com" }] }'

Chat with citations

/v1/kb/chattakes OpenAI-format messages, retrieves the most relevant chunks from the app's knowledge base, and answers grounded in them — the response includes the citations used. Use /v1/kb/search when you only want the matching chunks.

Shell

curl https://router-kb.intelli-verse-x.ai/v1/kb/chat \
  -H "Authorization: Bearer $INTELLIVERSE_API_KEY" \
  -d '{ "app_id": "APP_ID", "messages": [{"role":"user","content":"How do refunds work?"}] }'

Want a zero-code path? The MCP server exposes kb_ingest, kb_search, and kb_chat so your coding agent can manage knowledge bases conversationally.