Skip to main content
  1. Sign up — Create an account at platform.respan.ai
  2. Create an API key — Generate one on the API keys page
  3. Add credits or a provider key — Add credits on the Credits page or connect your own provider key on the Integrations page
Add the Docs MCP to your AI coding tool to get help building with Respan. No API key needed.
{
  "mcpServers": {
    "respan-docs": {
      "url": "https://respan.ai/docs/mcp"
    }
  }
}

​
What is Milvus?

Milvus is an open-source vector database built for scalable similarity search and AI applications. Milvus Lite provides a lightweight local version for development.

​
Setup

1

Install packages

pip install respan-tracing pymilvus openai
2

Set environment variables

export RESPAN_API_KEY="YOUR_RESPAN_API_KEY"
export OPENAI_API_KEY="YOUR_OPENAI_API_KEY"
3

Initialize Respan and query Milvus

Respan auto-instruments Milvus — search, insert, and delete operations are captured as spans.
from respan_tracing import RespanTelemetry
from respan_tracing.decorators import workflow, task
from pymilvus import MilvusClient
from openai import OpenAI

# Initialize — auto-instruments Milvus
telemetry = RespanTelemetry()
client = OpenAI()
milvus = MilvusClient("milvus_demo.db")  # Milvus Lite for local development

# Create collection
milvus.create_collection(collection_name="docs", dimension=1536)

docs = [
    "Respan provides observability for LLM applications.",
    "Traces capture the full lifecycle of an LLM request.",
]


def get_embeddings(texts: list[str]) -> list[list[float]]:
    response = client.embeddings.create(model="text-embedding-3-small", input=texts)
    return [item.embedding for item in response.data]


# Seed documents with embeddings
doc_embeddings = get_embeddings(docs)
milvus.insert(
    collection_name="docs",
    data=[
        {"id": i, "vector": doc_embeddings[i], "text": docs[i]}
        for i in range(len(docs))
    ],
)


@task(name="search_docs")
def search_docs(query: str):
    query_embedding = get_embeddings([query])
    results = milvus.search(
        collection_name="docs",
        data=query_embedding,
        limit=3,
        output_fields=["text"],
    )
    return [hit["entity"]["text"] for hit in results[0]]


@workflow(name="rag_pipeline")
def rag_pipeline(query: str):
    context = search_docs(query)
    response = client.chat.completions.create(
        model="gpt-4o-mini",
        messages=[
            {"role": "system", "content": f"Context: {chr(10).join(context)}"},
            {"role": "user", "content": query},
        ],
    )
    return response.choices[0].message.content


result = rag_pipeline("How does tracing work?")
print(result)
4

View your trace

Open the Traces page to see Milvus operations as spans in your trace tree.

​
Configuration

Milvus is auto-instrumented via Instruments.MILVUS. No additional configuration is needed.
Milvus tracing is currently available in the Python SDK only.
See the Python Tracing SDK reference for configuration options.