I’m trying to integrate mcp-atlassian with n8n, but I haven’t been able to get it working so far. I’ve successfully connected other MCPs without any issues, but this specific one is giving me trouble.
I’m running n8n locally using the official Docker image (n8nio/n8n). I’ve tried customizing the container to install mcp-atlassian, and although the installation seems to succeed, I can’t get n8n to recognize or communicate with it properly.
Interestingly, I was able to add and run mcp-atlassian without any problems using Cursor, so the MCP itself seems to work fine. However, I haven’t found a way to make it work within n8n.
Has anyone managed to get this working with n8n? If so, could you please share how you did it?
The n8n MCP client only supports the SSE method. You have to deploy a single server and run it in the following way to serve it, then set the server’s endpoint in the MCP client node to enable calls.
I found the solution! Just install phyton3 in the docker image. I attach the Dockerfile updated. Try it!
# I used version 1.86.1 because a bug I am waiting to be resolved
FROM n8nio/n8n:1.86.1
USER root
# Install necessary system packages
RUN apk add --no-cache \
curl \
git \
build-base \
chromium \
bash \
tar \
xz \
util-linux \
coreutils \
python3 \
py3-pip
# Install Astral uv/uvx and make available system-wide
RUN curl -Ls https://astral.sh/uv/install.sh | bash \
&& cp /root/.local/bin/uv /usr/local/bin/uv \
&& cp /root/.local/bin/uvx /usr/local/bin/uvx \
&& chmod a+rx /usr/local/bin/uv /usr/local/bin/uvx \
&& mkdir -p /data/mcp \
&& chown -R node:node /data/mcp
ENV PATH="/usr/local/bin:/root/.local/bin:${PATH}"
USER node
You were a savior. Your docker approach for invoking uvx worked. and its working fine.
However, i am facing some connectivity issues recently, it was all working fine sometime back but then its giving me timeout errors now. Do you know what could be the reason?