Как создать AI Agent с Agno и MCP-сервером ClickHouse

В этом руководстве вы узнаете, как создать AI Agent на базе Agno, который может взаимодействовать с Песочницей ClickHouse с помощью MCP-сервера ClickHouse.

Пример ноутбукаЭтот пример доступен в виде ноутбука в репозитории examples.

Предварительные требования

На вашей системе должен быть установлен Python.
На вашей системе должен быть установлен pip.
Вам понадобится API key Anthropic или API key другого провайдера LLM

Следующие шаги можно выполнить либо в Python REPL, либо с помощью скрипта.

Установка библиотек

Установите библиотеку Agno, выполнив следующие команды:

pip install -q --upgrade pip
pip install -q agno
pip install -q ipywidgets

Настройка учетных данных

Далее вам потребуется указать API-ключ Anthropic:

import os, getpass
os.environ["ANTHROPIC_API_KEY"] = getpass.getpass("Enter Anthropic API Key:")

Response

Enter Anthropic API Key: ········

Использование другого провайдера LLMЕсли у вас нет API-ключа Anthropic и вы хотите использовать другого провайдера LLM, инструкции по настройке учетных данных можно найти в документации Agno

Затем определите учетные данные, необходимые для подключения к Песочнице ClickHouse:

env = {
    "CLICKHOUSE_HOST": "sql-clickhouse.clickhouse.com",
    "CLICKHOUSE_PORT": "8443",
    "CLICKHOUSE_USER": "demo",
    "CLICKHOUSE_PASSWORD": "",
    "CLICKHOUSE_SECURE": "true"
}

Инициализируйте MCP-сервер ClickHouse и агент Agno

Теперь настройте MCP-сервер ClickHouse так, чтобы он указывал на Песочницу ClickHouse, а затем инициализируйте агент Agno и задайте ему вопрос:

from agno.agent import Agent
from agno.tools.mcp import MCPTools
from agno.models.anthropic import Claude

async with MCPTools(command="uv run --with mcp-clickhouse --python 3.13 mcp-clickhouse", env=env, timeout_seconds=60) as mcp_tools:
    agent = Agent(
        model=Claude(id="claude-3-5-sonnet-20240620"),
        markdown=True,
        tools = [mcp_tools]
    )
await agent.aprint_response("What's the most starred project in 2025?", stream=True)

Response

▰▱▱▱▱▱▱ Thinking...
┏━ Message ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃                                                                                                                 ┃
┃ What's the most starred project in 2025?                                                                        ┃
┃                                                                                                                 ┃
┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛
┏━ Tool Calls ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃                                                                                                                 ┃
┃ • list_tables(database=github, like=%)                                                                          ┃
┃ • run_select_query(query=SELECT                                                                                 ┃
┃     repo_name,                                                                                                  ┃
┃     SUM(count) AS stars_2025                                                                                    ┃
┃ FROM github.repo_events_per_day                                                                                 ┃
┃ WHERE event_type = 'WatchEvent'                                                                                 ┃
┃     AND created_at >= '2025-01-01'                                                                              ┃
┃     AND created_at < '2026-01-01'                                                                               ┃
┃ GROUP BY repo_name                                                                                              ┃
┃ ORDER BY stars_2025 DESC                                                                                        ┃
┃ LIMIT 1)                                                                                                        ┃
┃                                                                                                                 ┃
┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛
┏━ Response (34.9s) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃                                                                                                                 ┃
┃ To answer your question about the most starred project in 2025, I'll need to query the ClickHouse database.     ┃
┃ However, before I can do that, I need to gather some information and make sure we're looking at the right data. ┃
┃ Let me check the available databases and tables first.Thank you for providing the list of databases. I can see  ┃
┃ that there's a "github" database, which is likely to contain the information we're looking for. Let's check the ┃
┃ tables in this database.Now that we have information about the tables in the github database, we can query the  ┃
┃ relevant data to answer your question about the most starred project in 2025. We'll use the repo_events_per_day ┃
┃ table, which contains daily event counts for each repository, including star events (WatchEvents).              ┃
┃                                                                                                                 ┃
┃ Let's create a query to find the most starred project in 2025:Based on the query results, I can answer your     ┃
┃ question about the most starred project in 2025:                                                                ┃
┃                                                                                                                 ┃
┃ The most starred project in 2025 was deepseek-ai/DeepSeek-R1, which received 84,962 stars during that year.     ┃
┃                                                                                                                 ┃
┃ This project, DeepSeek-R1, appears to be an AI-related repository from the DeepSeek AI organization. It gained  ┃
┃ significant attention and popularity among the GitHub community in 2025, earning the highest number of stars    ┃
┃ for any project during that year.                                                                               ┃
┃                                                                                                                 ┃
┃ It's worth noting that this data is based on the GitHub events recorded in the database, and it represents the  ┃
┃ stars (WatchEvents) accumulated specifically during the year 2025. The total number of stars for this project   ┃
┃ might be higher if we consider its entire lifespan.                                                             ┃
┃                                                                                                                 ┃
┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛

​Предварительные требования

Предварительные требования