companyId?

string | undefined

(Optional) The unique identifier of the tenant when a service token is used.

maxOutputTokens?

number | undefined

(Optional) Maximum number of tokens in the generated answer (1-10000). Higher values allow longer, more detailed answers. Lower values force concise responses. Defaults to 1000 tokens. Note: higher values increase latency and cost

model?

string | undefined

(Optional) The specific provider and model identifier. Examples: 'openai://gpt-4o' or 'openai://gpt-4o-mini' for OpenAI, 'mistral://mistral-small-2506' for Mistral. Check provider documentation for available models

searchQuery

string

The question or query to be answered using the knowledge base. The LLM will search the knowledge base and generate an answer based on found results. Example: 'What is the product knowledge base?'

searchResults

(SearchKnowledgeBaseResultItem)[]

The search results to be used to generate the answer. The LLM will use the search results to generate an answer based on the found results.

systemPrompt?

string | undefined

(Optional) Optional system prompt to customize LLM behavior and context. Sets instructions for how the LLM should answer questions. Example: 'You are a helpful technical support assistant. Answer only based on the provided knowledge base. If information is not found, say "I don't have this information in the knowledge base."'

temperature?

number | undefined

(Optional) Controls answer randomness and creativity (0.0-1.0). Low values (0.0-0.3) produce consistent, deterministic answers. High values (0.7-1.0) produce more diverse and creative responses. Defaults to 0.3 for factual consistency