How to Use the Qdrant MCP in OpenAI Agents SDK

Q: Can an agent delete vector points using Qdrant and OpenAI Agents SDK?

Yes, the agent can call the delete tool using specific point IDs via the MCP interface. You should configure the OpenAI Agents SDK built-in guardrails to require a manual confirmation step before running this irreversible action on Qdrant. This prevents autonomous agents from clearing out vectors without developer oversight.

Q: How does the OpenAI Agents SDK handle Qdrant collection pagination?

The agent uses the scroll tool to paginate through Qdrant points and retrieve their associated payloads. By setting the limit and offset parameters, your OpenAI Agents SDK agent can systematically inspect entire collections without hitting memory limits. The OpenAI tracing dashboard logs each pagination step for debugging.

Q: What is the fastest way to verify a Qdrant collection exists?

Use the listcollections tool to get a full list of active collections in your Qdrant database. If the target collection is present, your OpenAI Agents SDK agent can then call getcollection to check its configuration and vector size. This workflow prevents runtime errors before you initiate similarity searches.

Run OpenAI Agents SDK with direct access to your Qdrant vector database for secure, production-grade similarity searches.

See Vinkius in Action

Works with every AI agent you already use

…and any MCP-compatible client

MCP Servers — Included with Plan

Connect Qdrant MCP to OpenAI Agents SDK

Create your Vinkius account to connect Qdrant to OpenAI Agents SDK — we handle the hosting, security, and runtime updates so you don't have to. No server setup required.

GDPR Included with Plan

Setup Qdrant with OpenAI Agents SDK

Ask AI about this MCP

ChatGPT

Claude

Perplexity

Safe vector search via OpenAI Agents SDK

The `search` tool lets your OpenAI Agents SDK agent query your Qdrant database using float arrays to find nearest neighbors. Your Python-based agent uses the SDK's runtime guardrails to validate the search payload before sending it via this MCP Server, stopping malformed float arrays before they hit your database. If the initial Qdrant vector search returns low-confidence scores, the OpenAI Agents SDK agent can automatically fall back to the `scroll` tool to paginate through collection points. It's an easy way to ensure your production system avoids silent failures by letting the OpenAI dashboard trace every database call.

Verified point deletion with built-in guardrails

The `delete` tool removes specific points from your Qdrant index to keep your vector database clean. Because this action is irreversible, the OpenAI Agents SDK built-in verification rules prompt for confirmation or check agent state before running the deletion. Before executing the Qdrant purge, your OpenAI Agents SDK agent can call `get_points` to verify the target IDs match the intended records. That's how you prevent autonomous agents from wiping out the wrong high-dimensional vectors during routine cleanup tasks.

Real-time audits using this MCP Server

The `count` tool returns the exact number of points in any given Qdrant collection to verify your vector density. Your OpenAI Agents SDK agent can run this check alongside `get_collection` to monitor index health and ensure data ingestion is actually working. The OpenAI Agents SDK agent uses multi-agent handoffs to pass these Qdrant metrics to a supervisor agent. If the point count drops below your threshold, the system triggers an alert, keeping your production vector pipelines operating within safe parameters.

Setup guide

Set up Qdrant MCP in OpenAI Agents SDK

Prerequisites

Python 3.10+ installed
openai-agents package (pip install openai-agents)
Active Vinkius subscription with a valid endpoint token

1

Install the SDK

Run pip install openai-agents to install the OpenAI Agents SDK. The MCP integration is built-in — no extra dependencies needed.
2

Connect via SSE transport

Use MCPServerSse with your Vinkius endpoint URL. Replace [YOUR_TOKEN_HERE] with your token from cloud.vinkius.com. The SDK auto-discovers all Qdrant tools at runtime.
3

Create your Agent

Pass the MCP to Agent(mcp_servers=[server]). The agent receives Qdrant tools as native definitions — JSON schemas resolve automatically.
4

Run the agent

Call Runner.run(agent, prompt) to execute. The agent invokes the appropriate Qdrant tools and returns structured results. Copy the full example on the right to get started.

agent.py

import asyncio
from agents import Agent, Runner
from agents.mcp import MCPServerSse

async def main():
    async with MCPServerSse(
        url="https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp"
    ) as server:
        agent = Agent(
            name="Qdrant Agent",
            instructions="You have access to Qdrant tools.",
            mcp_servers=[server],
        )
        result = await Runner.run(agent, "List recent transactions")
        print(result.final_output)

asyncio.run(main())

Get your connection token →

Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Qdrant. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.

Why Choose Vinkius

Vinkius connects your tools to AI with real-time monitoring and automatic cost savings — all from one dashboard.

Connect Qdrant now

Real-time monitoring

Live

visibility into every interaction

Connect your favorite tools to your AI and see exactly what's happening — every request, every response, in real time.

Built-in savings

60%

lower AI costs

Vinkius compresses data between your apps and your AI automatically. Lower bills every month — no configuration required.

Single dashboard

One

place for every integration

Every tool your AI connects to, managed from a single screen. One account, complete control.

Common questions about Qdrant MCP in OpenAI Agents SDK

You pass the query vector as a JSON array of floats directly to the `search` tool within your agent definition. The OpenAI Agents SDK validates the array dimensions at runtime before hitting your Qdrant database. This keeps your vector queries fast and prevents malformed vectors from causing connection timeouts.

Yes, the agent can call the `delete` tool using specific point IDs via the MCP interface. You should configure the OpenAI Agents SDK built-in guardrails to require a manual confirmation step before running this irreversible action on Qdrant. This prevents autonomous agents from clearing out vectors without developer oversight.

The agent uses the `scroll` tool to paginate through Qdrant points and retrieve their associated payloads. By setting the limit and offset parameters, your OpenAI Agents SDK agent can systematically inspect entire collections without hitting memory limits. The OpenAI tracing dashboard logs each pagination step for debugging.

Use the `list_collections` tool to get a full list of active collections in your Qdrant database. If the target collection is present, your OpenAI Agents SDK agent can then call `get_collection` to check its configuration and vector size. This workflow prevents runtime errors before you initiate similarity searches.

Your vector float arrays and payload metadata remain isolated within the Vinkius V8 sandbox during execution. The server only transmits the point IDs and query vectors directly to your specified Qdrant database endpoint. No raw data is cached or stored externally, keeping your proprietary vector embeddings secure.

Use it with your favorite AI tools

Connect this server to Cursor, Claude, VS Code, and more.

OpenAI Agents SDK sdk-python

Google ADK sdk-python

Pydantic AI sdk-python

Vercel AI SDK sdk-typescript