add glama.ai integrations (or biocontext.ai...)

[cmungall]

We want to make it easy for someone to add MCPs without faffing around and making a yaml file.

Let's say I find an MCP I want to use on @punkpeye's glama.ai, like for biomart:

https://glama.ai/mcp/servers/@jzinno/biomart-mcp

I want to do an immediate query like this

metacoder -c claude "what are the coordinates of tp53" -e glama.ai:jzinno/biomart-mcp

(exact slug string tbd)

right now it's more tedious, I or an AI need to navigate to the page find this

{
    "mcpServers": {
        "Biomart": {
            "command": "uv",
            "args": [
                "run",
                "--with",
                "mcp[cli]",
                "--with",
                "pybiomart",
                "mcp",
                "run",
                "/your/path/to/biomart-mcp.py"
            ]
        }
    }
}

and convert to the yaml format metacoder uses...

In theory metacoder should be able to get the metadata that includes the uv run command from the page

curl -X GET 'https://glama.ai/api/mcp/v1/servers/jzinno/biomart-mcp'

But this doesn't seem to be included yet...

[dragon-ai-agent]

Status Update on Glama.ai Integration

I investigated the current status of this feature request:

Current Challenge:
As noted in the issue, the glama.ai API endpoint (https://glama.ai/api/mcp/v1/servers/jzinno/biomart-mcp) does not currently include the necessary metadata like the uv run command needed for automatic configuration.

Research Needed:

1. Check if glama.ai has updated their API to include installation/execution metadata
2. Investigate if biocontext.ai or other MCP registries provide better programmatic access
3. Explore scraping the web interface as a fallback option

Implementation Approach:
If the metadata becomes available, the implementation would involve:

1. Adding a new extension syntax parser for glama.ai:username/package format
2. Creating an HTTP client to fetch MCP metadata from registry APIs
3. Converting the metadata to metacoder YAML format automatically
4. Caching the converted configurations locally

Status: Blocked pending API improvements from glama.ai or alternative registries.

Would it be valuable to explore other MCP registries that might already provide this metadata, or should we wait for glama.ai to enhance their API?

[punkpeye]

Just seeing this now

If you need help with Glama APIs etc, join https://glama.ai/discord and I will be happy to support.