Hey everyone,
I’m trying to set up Penpot Copilot using Docker Desktop and a local backend, but I’m running into connectivity issues between the two.
My Setup:
Penpot Frontend (UI):
I’m running the official Penpot frontend in Docker Desktop. I cloned the repo and use a Docker Compose file so that the UI is available at localhost:9001.
Penpot Copilot Backend:
I have the copilot backend running locally (outside of Docker) with Node.js on port 3000. My .env file in the backend folder looks like this:
OPENAI_KEY=local
MODEL_API_URL=http://localhost:11434/api/generate
PENPOT_DATABASE_URL=file:./penpot.db
This backend is supposed to send AI requests to my local DeepSeek instance.
DeepSeek via Ollama:
I’m using Ollama to run DeepSeek, which is available at localhost:11434/api/generate.
The Problem:
Even though I’ve configured my frontend (via the Docker Compose file or a .env file) to point to my backend at localhost:3000, the Penpot Copilot plugin doesn’t seem to be reliably connecting to it. I sometimes see HTTP errors that hint at protocol or connectivity issues, and I’m not 100% sure that the requests from the frontend are even reaching my backend.
I’ve checked Docker Desktop logs and confirmed that the frontend container is running, and my backend is up when I run it locally. However, there’s still a disconnect, and I can’t verify if the API calls are actually going to my local backend endpoints.
What I’m Looking For:
Debugging Help: How can I verify that the Penpot Copilot plugin in the frontend is correctly connecting to my local backend at localhost:3000?
Docker Desktop Tips: Are there any best practices or debugging steps (e.g., checking container logs, network traffic) that could help me determine if the requests from the plugin are hitting my backend?
Any advice or step-by-step instructions to troubleshoot this connectivity issue would be really appreciated!
Thanks in advance for your help.
