Settings

Tune retrieval and generation per knowledge base.

Settings control how documents are retrieved and how the model responds. Manage them from the dashboard or via API.

Get settings

GET/kbs/{kb_id}/settings

curl

curl https://api.docubix.com/kbs/{kb_id}/settings -H "X-Clerk-User-Id: your_clerk_user_id"

Update settings

PATCH/kbs/{kb_id}/settings

curl

curl -X PATCH https://api.docubix.com/kbs/{kb_id}/settings -H "X-Clerk-User-Id: your_clerk_user_id" \
  -H "Content-Type: application/json" \
  -d '{"temperature": 0.3, "retrieval_top_k": 8}'

Settings fields

Field	Description	Default
system_prompt	Instructions prepended to every answer	Educational assistant prompt
temperature	Model creativity (0.0–2.0)	0.2
model_name	Override LLM model (empty = server default)	Server default
retrieval_top_k	Chunks retrieved from vector search	5
max_context_chunks	Chunks sent to the LLM as context	5
fallback_response	Reply when no relevant context is found	Couldn't find message
rate_limit_per_minute	Optional per-KB rate limit	Not set

Default system prompt

text

You are an educational assistant.
Answer ONLY using the provided context.
If the answer cannot be found in the context, say:
"I couldn't find that information in the knowledge base."

Default fallback

text

I couldn't find that information in the knowledge base.