Documentation

Build on Startrag

Everything you need to turn flat document chunks into a navigable constellation of meaning.

Quickstart

Index your first documents and run a graph-aware query in under five minutes.

Concepts

Stars, clusters, and constellations — how Startrag models knowledge.

API reference

Endpoints for ingest, retrieve, and graph traversal.

Guides

Patterns for support copilots, agent memory, and RAG evaluation.

SDKs

Official clients for Python and TypeScript.

Security

Data handling, encryption, and compliance.

Quickstart

Install the SDK and index your first knowledge base.

pip install startrag

from startrag import Startrag

client = Startrag(api_key="sk_live_...")

# Ingest documents — they become "stars"
client.ingest(path="./docs", model="openai/gpt-5-mini")

# Retrieve with graph-aware context
answer = client.query(
    "How does constellation linking reduce hallucinations?"
)
print(answer.text)