Skip to main content
Glama
module-loading.md962 B
# Module Loading & Capability Aggregation Master MCP loads backend servers from multiple sources and aggregates their capabilities. ## Sources - `local`: A locally running server exposing HTTP endpoints - `git`, `npm`, `pypi`, `docker`: Stubs for different origins; endpoint resolution is config-driven (e.g., `config.port` or `url`). Example server block: ```yaml servers: - id: search type: local auth_strategy: master_oauth config: port: 4100 ``` ## Health Checks `DefaultModuleLoader` pings each server’s `/health` endpoint when loading to set an initial status (`running` or `error`). ## Capability Aggregation `CapabilityAggregator` discovers tools and resources via: - `GET /capabilities` (optional if provided by backend) - `POST /mcp/tools/list` - `POST /mcp/resources/list` Capabilities can be prefixed by server id (default) to avoid naming conflicts. Use the aggregated names in requests, e.g., `serverId.toolName`.

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/Jakedismo/master-mcp-server'

If you have feedback or need assistance with the MCP directory API, please join our Discord server