[rag.embeddings]
batch_size = 32
dimension = 384
model = "all-MiniLM-L6-v2"
provider = "local"

[rag.ingestion]
auto_ingest = true
chunk_size = 512
doc_types = [
    "md",
    "txt",
    "toml",
]
overlap = 50

[rag.llm]
api_url = "http://localhost:11434"
max_tokens = 2048
model = "llama3.2"
provider = "ollama"
temperature = 0.7

[rag.rag]
enabled = true

[rag.retrieval]
hybrid = false
reranking = false
similarity_threshold = 0.7
top_k = 5

[rag.vector_db]
db_type = "memory"
namespace = "provisioning-solo"