[BUG CLIENT]: Agent is not working with Remote MCP server

[JadyLiu]

Python -VV

Python 3.13.2 (main, Feb  4 2025, 14:51:09) [Clang 16.0.0 (clang-1600.0.26.6)]

Pip Freeze

annotated-types==0.7.0
anyio==4.9.0
Authlib==1.6.0
aws-requests-auth==0.4.3
boto3==1.38.23
botocore==1.38.23
certifi==2025.4.26
cffi==1.17.1
charset-normalizer==3.4.2
click==8.1.8
colorama==0.4.6
cryptography==45.0.3
Deprecated==1.2.18
dill==0.4.0
docstring-parser==0.15
eval_type_backport==0.2.2
exceptiongroup==1.3.0
fastapi==0.115.12
fastapi-mcp==0.3.4
fastmcp==2.5.2
googleapis-common-protos==1.70.0
griffe==1.7.3
h11==0.16.0
httpcore==1.0.9
httpx==0.28.1
httpx-sse==0.4.0
idna==3.10
importlib_metadata==8.6.1
itsdangerous==2.2.0
Jinja2==3.1.6
jmespath==1.0.1
Markdown==3.8
markdown-it-py==3.0.0
MarkupSafe==3.0.2
mcp==1.9.1
mdurl==0.1.2
mistralai==1.8.1
mpmath==1.3.0
openapi-pydantic==0.5.1
opentelemetry-api==1.33.1
opentelemetry-exporter-otlp-proto-common==1.33.1
opentelemetry-exporter-otlp-proto-http==1.33.1
opentelemetry-proto==1.33.1
opentelemetry-sdk==1.33.1
opentelemetry-semantic-conventions==0.54b1
pillow==11.2.1
prompt_toolkit==3.0.51
protobuf==5.29.4
pycparser==2.22
pydantic==2.11.5
pydantic-settings==2.9.1
pydantic_core==2.33.2
Pygments==2.19.1
PyJWT==2.10.1
python-dateutil==2.9.0.post0
python-dotenv==1.1.0
python-multipart==0.0.20
requests==2.32.3
rich==14.0.0
s3transfer==0.13.0
shellingham==1.5.4
six==1.17.0
slack_bolt==1.23.0
slack_sdk==3.35.0
sniffio==1.3.1
sse-starlette==2.3.5
starlette==0.46.2
strands-agents==0.1.4
strands-agents-tools==0.1.3
sympy==1.14.0
tenacity==9.1.2
tomli==2.2.1
typer==0.15.4
typing-inspection==0.4.1
typing_extensions==4.13.2
urllib3==2.4.0
uvicorn==0.34.2
watchdog==6.0.0
wcwidth==0.2.13
websockets==15.0.1
wrapt==1.17.2
zipp==3.21.0

Reproduction Steps

1. We have an existing remote MCP server working with current MCP client.
2. We try to switch to Mistral agent by following example: https://docs.mistral.ai/agents/mcp/
3. It works when using async with RunContext(model=MODEL,) as run_ctx: it does not work when using async with RunContext(agent_id=self.agent.id,) as run_ctx:
4. The output is
   I'm sorry, but I currently don't have the tools needed to process and summarize the loan application for Jane Doe. If you have any other questions or need further assistance, feel free to ask!

Expected Behavior

The Agent is able to connect to remote MCP server using SSE.

Additional Context

The client code snippet is below:

#!/usr/bin/env python
import asyncio
import os

from mistralai import Mistral
from mistralai.extra.run.context import RunContext
from mistralai.extra.mcp.sse import MCPClientSSE, SSEServerParams
from pathlib import Path

# Set the current working directory and model to use
cwd = Path(__file__).parent
MODEL = "mistral-medium-latest"

# Define the URL for the remote MCP server
server_url = "http://myworkingserver:8000/mcp"
mcp_client = MCPClientSSE(sse_params=SSEServerParams(url=server_url, timeout=100))
query = (
    "Summarize the loan application for "
    '{"name": "Jane Doe","age": 35,"income": 2000000,'
    '"loan_amount": 4500000,"credit_score": 820,'
    '"existing_liabilities": 15000,"purpose": "Home Renovation"}'
)

class ClaimAgent:

    def __init__(self):
        self.client = None
        self.agent = None

    async def initialize(self):
        # Initialize the Mistral client with your API key
        api_key = os.environ["MISTRAL_API_KEY"]
        self.client = Mistral(api_key)

        self.agent = self.client.beta.agents.create(
            model=MODEL,
            name="claim-agent",
            instructions="Agent to answer user queries and response the loan summary using available tools.",
            description="Loan Summary Agent",
        )

    async def process_claim(self, query: str):
        """Process user query using the agent with MCP servers"""

        if not self.client or not self.agent:
            await self.initialize()
    
        # Create a run context for the agent
        async with RunContext(
            agent_id=self.agent.id,
            continue_on_fn_error=False,
        ) as run_ctx:
            # Register the MCP client with the run context
            await run_ctx.register_mcp_client(mcp_client=mcp_client)

            # Run the agent with a query
            print("Running the agent with the query...")
            run_result = await self.client.beta.conversations.run_async(
                run_ctx=run_ctx,
                inputs=query,
            )

            print("All run entries:")
            for entry in run_result.output_entries:
                print(f"{entry}")
                print()
            print(f"Final Response: {run_result.output_as_text}")



if __name__ == "__main__":
    claim_agent = ClaimAgent()
    try:
        asyncio.run(claim_agent.process_claim(query))
    except Exception as e:
        print(f"❌ Exception occurred: {e}")

Suggested Solutions

No response

[aac228]

Hey @JadyLiu, thanks for creating the issue!

I've run the code and it was working on my own MCP server. I would propose two things:

• First you could try another model like mistral-large-latest to see if it performs better on your use case.
• Otherwise to understand better what could be wrong, I would add some debugging information. Could you add the following to your code:

if __name__ == "__main__":
    logging.basicConfig(level=logging.INFO)
    ....

That will show if the tools from your MCP server are properly added to the context, you should get logs like this:

INFO:mistralai.extra.run.context:Adding tool <tool-name> from MCPClientSSE

[JadyLiu]

Thank you @aac228! Confirm that it works when switching to mistral-large-latest model.