RAG

Memonic’s RAG pipeline lets you ingest documents and retrieve relevant chunks for any prompt — grounding your agent in up-to-date knowledge.

Ingest a document

curl -X POST https://memonic.dev/api/rag/ingest \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "documentId": "doc-runbook-001",
    "content": "To restart the API: run npm run deploy from the worker directory. To check live logs: use the dashboard at app.memonic.dev.",
    "metadata": {"source": "runbook", "version": "v2"}
  }'

Memonic chunks the document, generates embeddings via Workers AI (bge-base-en-v1.5), and stores them in Vectorize.

Query for relevant chunks

curl -X POST https://memonic.dev/api/rag/query \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "how do I restart the worker",
    "topK": 3,
    "filter": {"source": "runbook"}
  }'

Returns the most semantically relevant chunks to inject into your prompt context.