Optional
BM25_The average length of the chunks in the index for BM25
Optional
BM25_The BM25 B parameter
Optional
BM25_Whether to use BM25
Optional
BM25_The BM25 K parameter
Optional
DISABLE_Whether to disable analytics
Optional
DISTANCE_Optional
EMBEDDING_The base URL for the embedding API
Optional
EMBEDDING_The name of the embedding model to use
Optional
EMBEDDING_The prefix to use for the embedding query
Optional
EMBEDDING_The size of the embeddings
Optional
FREQUENCY_The frequency penalty to use
Optional
FULLTEXT_Whether to use fulltext search
Optional
INDEXED_Whether to only use indexed chunks
Optional
LLM_The base URL for the LLM API
Optional
LLM_The default model to use for the LLM
Optional
LOCKED?: boolean | nullWhether the dataset is locked to prevent changes or deletion
Optional
MAX_The maximum limit for the number of chunks for counting
Optional
MAX_The maximum number of tokens to use in LLM Response
Optional
MESSAGE_The prompt to use for converting a message to a query
Optional
N_The number of retrievals to include with the RAG model
Optional
PAGEFIND_Whether to enable pagefind indexing
Optional
PRESENCE_The presence penalty to use
Optional
PUBLIC_Optional
QDRANT_Whether or not to insert chunks into Postgres
Optional
RAG_The prompt to use for the RAG model
Optional
RERANKER_The base URL for the reranker API
Optional
RERANKER_The model name for the Reranker API
Optional
SEMANTIC_Whether to use semantic search
Optional
STOP_The stop tokens to use
Optional
SYSTEM_The system prompt to use for the LLM
Optional
TEMPERATURE?: number | nullThe temperature to use
Optional
USE_Whether to use the message to query prompt
Lets you specify the configuration for a dataset