Agentic Tools

Qdrant provides tools and resources that help AI agents work with Qdrant more effectively. Use them to give your agent task-specific guidance, troubleshoot existing deployments, design and evaluate search systems, and apply Qdrant best practices to your specific use case.

Getting Started with AI Agents

qdrant.tech/llms.txt is the single entry point for AI agents. We’ve curated it to be under 30KB and to provide your agent with everything it needs to get started, manage, and troubleshoot your Qdrant search engine. It tells an agent what Qdrant is, how to set it up, how to use it correctly, and where to find everything else on qdrant.tech.

To get started, give this to your agent:

Agent Prompt Start building on Qdrant
set up qdrant with https://qdrant.tech/llms.txt

By default, it recommends a free Qdrant Cloud cluster, where nothing needs to run on your machine and you don’t need a credit card. This is our simplest setup and gives your agent more control with the Qdrant CLI. Additionally, Cloud Inference creates embeddings for free with the free Qdrant-hosted models, so you don’t need an embedding library or an embedding provider key.

The agent asks before it creates anything, and you never paste a key into the chat.

To run Qdrant on your own machine instead, just say so in your prompt. The agent will set up Qdrant in Docker, and FastEmbed to generate embeddings on your machine.

What Our llms.txt Contains

The file has a hand-written brief, followed by an index of the site that is generated on every build.

PartWhat it gives your agent
OverviewWhat Qdrant is, when to use it, and why to start on Qdrant Cloud. Self-hosted Qdrant is set up only when you ask for it.
SetupThe steps to a working Qdrant Cloud cluster: sign up, choose a cluster, connect through the Qdrant Cloud CLI or the Cloud Console, store keys in a .env file, and test the connection. Coding agents recommend the CLI and install it only after you agree.
MCP serversWhen to connect the Qdrant MCP servers, and how.
Core usageConnection details, a Python example that creates a collection and searches it with Cloud Inference, the Query API for hybrid search, and common qcloud commands.
RulesGuardrails that prevent common mistakes, such as filtering on fields without a payload index, choosing the wrong vector size or distance metric, and building multitenancy without access control. The agent also confirms with you before it takes an action that costs money or deletes data.
Deep guidanceHow to search the agent skills for a specific problem, with or without the Qdrant Advisor installed.
Site indexThe main pages of the Develop, Deploy, Ecosystem, and Learn parts of the documentation, the articles, and pricing, each with a one-line summary. Each part also links to a full index of every page in it.
Optionalllms-full.txt, which holds the brief and every page index in one file, plus the agent skills page, the API reference, and legal and security information.

Every link in the file points to the Markdown version of a page, which agents read more reliably than HTML. The file changes with the product, so agents should fetch it fresh rather than rely on a saved copy.

Agent Skills

Agent skills encode solutions architect knowledge as structured files your agent loads on demand. They’re organized around symptoms, so they answer “why is this happening, and what should I check first?” Reach for skills when something is already running and behaving unexpectedly: slow search, growing memory, a stuck optimizer, poor relevance.

When you set up through the CLI, llms.txt also installs the Qdrant Advisor, a skill that searches skills.qdrant.tech and loads only the guidance that matches your problem.

Prompts in the Docs

Short, ready-to-use prompts appear directly on the documentation pages where they are relevant, so you can use the prompts when you need them instead of looking for them elsewhere. Each prompt points to the agent skill your assistant should load. The prompt gets the task started; the skill provides the deeper Qdrant-specific guidance. The prompt index lists all available prompts.

Where llms.txt gets you started, a prompt in the docs is for a specific task once you’re building, such as designing a collection or setting up monitoring.

These are designed to help as an optional additional resource when humans are navigating Qdrant documentation themselves.

The CLI and MCP Servers

Skills and prompts tell your agent what to do. The Qdrant Cloud CLI and the Qdrant MCP servers let it act.

ToolWhat your agent can do with itWorks with
Qdrant Cloud CLI (qcloud)Create and manage clusters, database API keys, and backups from the command line. You log in once in your browser with qcloud auth login.Qdrant Cloud
Qdrant Cloud MCPManage clusters, serverless spaces, backups, access, and billing through tool calls. You sign in with OAuth. It also searches the documentation.Qdrant Cloud
Qdrant Docs MCPSearch the documentation, articles, blog, and courses, and find code snippets. No account needed.Any deployment
Qdrant MCP (open source)Store and retrieve information in a Qdrant collection, as semantic memory or code search.Any deployment, including local

The CLI suits coding agents that can run shell commands, such as Claude Code, Cursor, or Codex. It is also the only one of these tools that creates database API keys. The Qdrant Cloud MCP server suits assistants that can’t run commands, such as claude.ai and Claude Desktop. If you run open source Qdrant on your own machine or servers, the CLI and the Qdrant Cloud MCP server don’t apply: use the Docs MCP server and the open source MCP server.

Was this page useful?

Thank you for your feedback! 🙏

We are sorry to hear that. 😔 You can edit this page on GitHub, or create a GitHub issue.