# smithery.yaml
# This file tells Smithery.ai how to build and run your agent.
build:
# Use the Dockerfile in the root of the repository
dockerfile: Dockerfile
startCommand:
type: http
configSchema:
type: object
properties:
session_timeout:
type: integer
description: "Session timeout in seconds"
default: 36000
max_sessions:
type: integer
description: "Maximum number of concurrent sessions"
default: 1000
required: []
run:
# Command to run the agent.
command: ["python", "-u", "server.py"] # -u flag for unbuffered output
# Ports to expose from the container.
ports:
- "8000:8000"
# Environment variables to set for the running agent.
environment:
PYTHONUNBUFFERED: "1" # Ensure Python output is not buffered
PYTHONDONTWRITEBYTECODE: "1" # Don't write .pyc files
MCP_SESSION_TIMEOUT: "360000" # Session timeout in seconds
MCP_MAX_SESSIONS: "1000" # Maximum concurrent sessions
MCP_STREAM_TIMEOUT: "30000" # Stream timeout in seconds
MCP_ENABLE_SSE: "true" # Enable Server-Sent Events
# Resource limits
resources:
memory: "512Mi" # Memory limit
cpu: "0.5" # CPU limit (50% of one core)
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/atimevil/dreamhack_mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server