Microsoft Agent Framework

Build a Microsoft Agent Framework agent that can search tweets, hand off IDs and cursors, post tweets, replay stored monitor events, and run extraction jobs - all through Xquik’s MCP server.

Prerequisites

Python 3.10+
Xquik API key (xq_...)
An LLM API key (OpenAI, Azure OpenAI, or any supported provider)

Install

pip install agent-framework

The agent-framework meta-package installs the core framework, providers, and MCP support.

Full Example

import asyncio
import os
from pathlib import Path
from agent_framework import Agent, MCPStreamableHTTPTool
from agent_framework.openai import OpenAIChatClient


async def main():
    async with Agent(
        client=OpenAIChatClient(model="gpt-4o"),
        name="xquik_agent",
        instructions="You help users interact with X through the Xquik API.",
        tools=MCPStreamableHTTPTool(
            name="xquik",
            url="https://xquik.com/mcp",
            description="X data platform with 118 MCP operations",
            header_provider=lambda kwargs: {
                "x-api-key": kwargs["xquik_api_key"],
            },
        ),
    ) as agent:
        handoff_prompt = (
            "Search for tweets about AI agents. Return compact JSON with "
            "query, route_used, tweets[{tweet_id,text,author_username,created_at}], "
            "has_more, next_cursor."
        )
        response = await agent.run(
            handoff_prompt,
            function_invocation_kwargs={
                "xquik_api_key": os.environ["XQUIK_API_KEY"],
            },
        )
        Path("xquik-agent-handoff.json").write_text(response.text, encoding="utf-8")


asyncio.run(main())

The agent discovers explore and xquik, then can call 118 MCP operations. The MCP runtime returns normalized snake_case fields through xquik.request(), so keep prompts aligned with tweet_id, has_more, next_cursor, and returned job IDs.

Handoff Checklist

Tweet search rows

Store tweet_id, text, author_username, created_at, has_more, next_cursor, and the original q.

User profile rows

Store source id as user_id, plus username, name, followers, verified, profile_picture, has_more, next_cursor, and the source lookup or search query.

Trend rows

Store each trend name, rank, query, and description. Keep response count, woeid, and the requested region with the run checkpoint.

Monitor and webhook setup

Store the returned monitor id as monitor_id, event_types, next_billing_at, the returned webhook id as webhook_id, url, and the one-time secret in a secret manager. On production deliveries, store delivery_id for receiver retry de-dupe and stream_event_id when one monitor event should process once across endpoint changes.

Stored event replay

Store event_id, type, monitor_id, monitor_type, occurred_at, has_more, next_cursor, and the after query for the next page.

Extraction jobs

Store extraction_id, status, poll, and export_after_complete; poll before loading CSV, JSON, or XLSX rows.

Writes

Store tweet_id or write_action_id, reply_to_tweet_id, status, charged_credits, and poll; do not resend pending writes.

Media attachments

For tweets or replies, pass public URLs in media and store tweet_id or write_action_id. For DMs, upload first, pass one media_id in media_ids, store message_id, and leave reply_to_message_id unset.

Streaming Responses

Pass stream=True to run for real-time token streaming:

async with agent:
    async for update in agent.run(
        "What are the trending topics on X right now?",
        stream=True,
        function_invocation_kwargs={"xquik_api_key": xquik_api_key},
    ):
        if update.text:
            print(update.text, end="")

Multi-Agent Handoff

Keep data access in one agent, then pass its compact result to a separate analysis agent. This makes the MCP credential boundary explicit.

import asyncio
import os
from pathlib import Path
from agent_framework import Agent, MCPStreamableHTTPTool
from agent_framework.openai import OpenAIChatClient


async def main():
    client = OpenAIChatClient(model="gpt-4o")
    researcher = Agent(
        client=client,
        name="researcher",
        instructions=(
            "Search X. Return compact JSON with "
            "tweet_id, author_username, text, created_at, has_more, next_cursor, "
            "and route_used."
        ),
        tools=MCPStreamableHTTPTool(
            name="xquik",
            url="https://xquik.com/mcp",
            header_provider=lambda kwargs: {
                "x-api-key": kwargs["xquik_api_key"],
            },
        ),
    )
    analyst = Agent(
        client=client,
        name="analyst",
        instructions="Analyze supplied X data. Do not fetch additional data.",
    )

    async with researcher, analyst:
        research = await researcher.run(
            "Find the top 10 most engaged tweets about AI agents.",
            function_invocation_kwargs={
                "xquik_api_key": os.environ["XQUIK_API_KEY"],
            },
        )
        result = await analyst.run(
            "Analyze sentiment and key themes in this JSON:\n" + research.text
        )
        Path("xquik-agent-workflow-handoff.json").write_text(
            result.text,
            encoding="utf-8",
        )


asyncio.run(main())

Per-Run Headers

Pass API keys through header_provider and function_invocation_kwargs for multi-tenant apps. The key remains scoped to the current run.

mcp_tool = MCPStreamableHTTPTool(
    name="xquik",
    url="https://xquik.com/mcp",
    description="X data platform",
    header_provider=lambda kwargs: {
        "x-api-key": kwargs["xquik_api_key"],
    },
)

async with Agent(
    client=OpenAIChatClient(model="gpt-4o"),
    name="xquik_agent",
    instructions="You help users interact with X through the Xquik API.",
    tools=mcp_tool,
) as agent:
    response = await agent.run(
        "Search for tweets about AI agents",
        function_invocation_kwargs={"xquik_api_key": user_api_key},
    )

Do not close over a shared user’s key in header_provider. Resolve it from the current call’s function_invocation_kwargs.

Azure OpenAI

Use the same OpenAIChatClient with explicit Azure routing inputs:

import os

from agent_framework import Agent
from agent_framework.openai import OpenAIChatClient

client = OpenAIChatClient(
    model=os.environ["AZURE_OPENAI_CHAT_MODEL"],
    azure_endpoint=os.environ["AZURE_OPENAI_ENDPOINT"],
    api_version=os.environ["AZURE_OPENAI_API_VERSION"],
    api_key=os.environ["AZURE_OPENAI_API_KEY"],
)

agent = Agent(
    client=client,
    name="xquik_agent",
    instructions="You help users interact with X through the Xquik API.",
    tools=mcp_tool,
)

Environment Variables

Store your API keys in a .env file instead of hardcoding them:

.env

XQUIK_API_KEY=xq_YOUR_KEY_HERE
OPENAI_API_KEY=sk-...

import os
from dotenv import load_dotenv
from agent_framework import Agent, MCPStreamableHTTPTool
from agent_framework.openai import OpenAIChatClient

load_dotenv()

mcp_tool = MCPStreamableHTTPTool(
    name="xquik",
    url="https://xquik.com/mcp",
    header_provider=lambda kwargs: {
        "x-api-key": kwargs["xquik_api_key"],
    },
    description="X data platform",
)

agent = Agent(
    client=OpenAIChatClient(model="gpt-4o"),
    tools=mcp_tool,
)

Package Versions

Package	Version
`agent-framework`	1.11.0+
`mcp`	1.24.0+

Getting Started

SDKs

Guides

Webhooks

MCP Server

Agent Plugins

Framework Guides

OAuth

MPP

Resources

Microsoft Agent Framework

Prerequisites

Install

Full Example

Handoff Checklist

Tweet search rows

User profile rows

Trend rows

Monitor and webhook setup

Stored event replay

Extraction jobs

Writes

Media attachments

Streaming Responses

Multi-Agent Handoff

Per-Run Headers

Azure OpenAI

Environment Variables

Package Versions

​Prerequisites

​Install

​Full Example

​Handoff Checklist

Tweet search rows

User profile rows

Trend rows

Monitor and webhook setup

Stored event replay

Extraction jobs

Writes

Media attachments

​Streaming Responses

​Multi-Agent Handoff

​Per-Run Headers

​Azure OpenAI

​Environment Variables

​Package Versions

Prerequisites

Install

Full Example

Handoff Checklist

Streaming Responses

Multi-Agent Handoff

Per-Run Headers

Azure OpenAI

Environment Variables

Package Versions