Connect to Enjab RAG
Wire an existing tool or agent to Enjab RAG, the shared knowledge base, and keep it current.
Every Enjab agent should be connected to Enjab RAG so it grounds its answers on the same shared operational knowledge instead of its own stale copy. If your tool is not connected yet, connect it. When Enjab RAG changes, the Enjab RAG changelog lists what changed and whether any action is needed (most changes are server-side and need none).
The update prompt
Hand an agent this message:
Connect this project to Enjab RAG, the shared knowledge base for Enjab's agents, and bring it
up to the current contract. Read https://developers.enjab.ae/llms.txt and the Enjab RAG docs
at https://developers.enjab.ae/docs/enjab-rag (plus the changelog at
https://developers.enjab.ae/docs/changelog/enjab-rag).
If this is an AGENT: add Enjab RAG as an MCP server at https://rag.enjab.ae/api/mcp
(transport: http) with the header "Authorization: Bearer <key>" (an enjabrag_... API key),
and add to the system prompt: "Before answering any question about how the clinics operate,
call the Enjab RAG `search` tool first and ground the answer in the results."
If this is a BACKEND/tool: call POST https://rag.enjab.ae/api/search with a read key to
retrieve, and POST /api/ingest with a write key to add knowledge.
Apply ONLY the changelog items this project has not adopted yet. Do not refactor anything
else, keep the change surface minimal.The agent fetches the docs, connects (or re-checks) the integration, and applies only the delta.
What to verify
The connection exists
Agents have Enjab RAG registered as an MCP server (https://rag.enjab.ae/api/mcp); backends call /api/search (and /api/ingest if they add knowledge). See Connect an agent and the REST API.
The key scope is right
A read key is enough to search and fetch; only ingestion jobs need a write key. Give each agent its own named key, minted at rag.enjab.ae/dash/keys. See Authentication.
Agents search before answering
The system prompt instructs the agent to call search before answering operational questions, so it grounds on shared knowledge rather than guessing.
Latest change
Enjab RAG launched (11 Jun 2026) with structure-aware + Contextual Retrieval chunking. Retrieval improves server-side with no action needed; re-ingest a document only if you want it re-chunked. See the changelog.