Hello, i deployed Acheron on a droplet using DigitalOcean, and been trying to configure it, and I get the following issues, I was able to configure supabase and chat LLM model and also crawl the PydenticAI docs, and it all works fine when using the acheron chat UI,however, when setting up the MCP server in Windsurf if a get a request failed error. in the MCP config file i set up the GRAPH_SERVICE_URL to point to: http://:8100, when i test it on the machine console i get
curl -H “Content-Type: application/json” http://localhost:8100
{“detail”:“Not Found”}
and when using the proser i get the same error, which tells me its working but its not correct URL? as nothing is returned.
I was able to solve the issue, and i can get to the agent API using curl, however, i now realize that the MCP configuration for docker requires the docker to run locally, and there is no MCP configuration to allow windsurf communicate with Acheron if deployed on a cloud instance. am i correct?
1 Like
I tried to use deploy inside Streamlit to deploy my agent code to Streamlit cloud but get the following message. Is there someway to configure?
That is correct! Right now Anthropic doesn’t have support for remote MCP servers, so you’ll have to have Archon running as an MCP server in the same machine you are using Windsurf.
Is your instance of Archon not within a GitHub repository on your local machine? I manually deploy my Streamlit pages in the cloud instead of using their deploy option, so I’m not an expert at this. That’s just the first thing that sticks out to me if it’s asking you to put it in a GitHub repo
I didn’t realize the deploy was a core Streamlit feature. I thought it was part of Archon. I’m good.
1 Like
Okay gotcha! Yeah I see the confusion - the deploy button is there in any Streamlit UI