SwarmsSwarms Cloud

MCP

A hosted Model Context Protocol server for the Swarms API. Point any MCP client at one URL to give it agents, swarms, and batch execution as callable tools. No local process to run.

Endpoint
https://mcp.swarms.world/mcp
Checking

Status

Checking…

Probing the endpoint from our region.

Handshake latency

—

Measured server-side from our region.

Observed uptime

—

No observations recorded yet on this server.

Incident history

Connect

The server speaks streamable HTTP and authenticates with the same x-api-key header as the REST API. Create a key in API keys and export it as SWARMS_API_KEY.

# pip install mcpimport asyncioimport osfrom mcp import ClientSessionfrom mcp.client.streamable_http import streamablehttp_clientSWARMS_MCP_URL = "https://mcp.swarms.world/mcp"async def main() -> None:    async with streamablehttp_client(        SWARMS_MCP_URL,        headers={"x-api-key": os.environ["SWARMS_API_KEY"]},    ) as (read, write, _):        async with ClientSession(read, write) as session:            await session.initialize()            # Every tool the server exposes, with its input schema.            tools = await session.list_tools()            for tool in tools.tools:                print(tool.name)asyncio.run(main())

Tools

Reading the tool list from the server…

Run a swarm

Call the swarm completions tool through the session you just opened. This example runs a ConcurrentWorkflow so both analysts work the same task in parallel and return together. Tool names come from list_tools, so run the connect snippet first to confirm what your key can reach.

import asyncioimport jsonimport osfrom mcp import ClientSessionfrom mcp.client.streamable_http import streamablehttp_clientSWARMS_MCP_URL = "https://mcp.swarms.world/mcp"async def main() -> None:    async with streamablehttp_client(        SWARMS_MCP_URL,        headers={"x-api-key": os.environ["SWARMS_API_KEY"]},    ) as (read, write, _):        async with ClientSession(read, write) as session:            await session.initialize()            result = await session.call_tool(                "run_swarm_v1_swarm_completions_post",                {                    "name": "Market Research Swarm",                    "description": "Three analysts research the same task in parallel",                    "swarm_type": "ConcurrentWorkflow",                    "task": "Analyze the impact of AI agents on modern healthcare",                    "agents": [                        {                            "agent_name": "Market Analyst",                            "system_prompt": "You analyze market trends and opportunities.",                            "model_name": "gpt-5.4",                            "max_loops": 1,                        },                        {                            "agent_name": "Risk Analyst",                            "system_prompt": "You identify risks and regulatory constraints.",                            "model_name": "claude-haiku-4-5",                            "max_loops": 1,                        },                    ],                    "max_loops": 1,                },            )            for block in result.content:                if block.type == "text":                    print(json.dumps(json.loads(block.text), indent=2))asyncio.run(main())

The same swarm without MCP, straight against the REST endpoint:

Swarm completion over HTTP

import osfrom typing import Any, TypedDictimport requestsclass Agent(TypedDict):    agent_name: str    system_prompt: str    model_name: str    max_loops: intclass SwarmCompletionRequest(TypedDict):    name: str    description: str    swarm_type: str    task: str    agents: list[Agent]    max_loops: intpayload: SwarmCompletionRequest = {    "name": "Market Research Swarm",    "description": "Three analysts research the same task in parallel",    "swarm_type": "ConcurrentWorkflow",    "task": "Analyze the impact of AI agents on modern healthcare",    "agents": [        {            "agent_name": "Market Analyst",            "system_prompt": "You analyze market trends and opportunities.",            "model_name": "gpt-5.4",            "max_loops": 1,        },        {            "agent_name": "Risk Analyst",            "system_prompt": "You identify risks and regulatory constraints.",            "model_name": "claude-haiku-4-5",            "max_loops": 1,        },    ],    "max_loops": 1,}response: requests.Response = requests.post(    "https://api.swarms.world/v1/swarm/completions",    headers={        "x-api-key": os.environ["SWARMS_API_KEY"],        "Content-Type": "application/json",    },    json=payload,)response.raise_for_status()data: dict[str, Any] = response.json()print(data)

Tutorials

TypeScript

Build an Agent CLI over MCP

Wire the MCP server into a command-line tool that runs agents from your terminal.

Read tutorialRust

Build a Multi-Agent Research Tool over MCP

Drive a multi-agent research swarm from a native Rust client.

Read tutorialPython

Run a Batch Pipeline over MCP

Fan a batch of tasks across agents and collect every result in one pipeline.

Read tutorial

Documentation

MCP client reference

Transport, authentication, and every tool the server exposes.

docs.swarms.ai/llms.txt

The full documentation index in one plaintext file, for agents that crawl it directly.