CLI

ragwise ships two CLI commands: ragwise init and ragwise serve.

ragwise init

Generates a fully-typed config file in the current directory.

ragwise init

Creates ragwise_config.py:

# ragwise_config.py — generated by `ragwise init`
# Edit and import into your project.

from ragwise import RAG, QueryConfig

rag = RAG(
    # LLM provider — change to "anthropic/claude-haiku-4-5" or "ollama/llama3"
    llm="openai/gpt-4o-mini",

    # Embedder — change to "local/all-MiniLM-L6-v2" for offline use
    embedder="openai/text-embedding-3-small",

    # Store — upgrade path:
    #   "memory"                         dev / CI (volatile)
    #   "lance://./ragwise-index"           persistent dev (no server)
    #   "postgresql://user:pass@host/db" production
    store="memory",

    # Chunking
    chunk_size=512,
    chunk_overlap=64,

    # LLM response caching
    cache=True,
)

query_config = QueryConfig(
    top_k=5,
    include_citations=True,
    check_sufficiency=True,
)

Import and use directly:

from ragwise_config import rag, query_config

async def main():
    async with rag:
        await rag.ingest("./docs/")
        answer = await rag.query("What is the refund policy?", config=query_config)
        print(answer.text)

ragwise serve

Starts a minimal HTTP server exposing the RAG pipeline as an API.

pip install ragwise[serve]
ragwise serve --store lance://./ragwise-index --llm openai/gpt-4o-mini --port 8080

Options:

Flag	Default	Description
--store	memory	Store backend (same strings as RAG(store=...))
--llm	openai/gpt-4o-mini	LLM provider string
--embedder	openai/text-embedding-3-small	Embedder string
--port	8080	Port to listen on
--host	0.0.0.0	Host to bind to

Endpoints

POST /ingest

curl -X POST http://localhost:8080/ingest \
  -H "Content-Type: application/json" \
  -d '{"path": "./docs/"}'

Response:

{"succeeded": 12, "failed": 0, "skipped": 0, "errors": []}

POST /query

curl -X POST http://localhost:8080/query \
  -H "Content-Type: application/json" \
  -d '{"question": "What is the refund policy?", "top_k": 5}'

Response:

{
  "text": "You can get a full refund within 30 days.",
  "citations": [
    {"source": "docs/policies.md", "text": "Our refund policy..."}
  ],
  "sufficient": true
}

GET /health

curl http://localhost:8080/health
# {"status": "ok"}

Docker

FROM python:3.11-slim
WORKDIR /app
COPY . .
RUN pip install ragwise[serve,lance]
CMD ["ragwise", "serve", "--store", "lance://./ragwise-index", "--port", "8080"]

docker build -t my-rag .
docker run -p 8080:8080 -e OPENAI_API_KEY=sk-... my-rag