Granoflow MCP server
{
"mcpServers": {
"granoflow": {
"command": "npx",
"args": [
"-y",
"@granoflow/mcp-server@latest"
],
"env": {
"GRANOFLOW_MCP_CONFIG_PATH": "/path/to/directory",
"GRANOFLOW_API_TOKEN": "<granoflow_api_token>",
"GRANOFLOW_API_BASE_URL": "https://..."
}
}
}
}{
"mcpServers": {
"granoflow": {
"command": "npx",
"args": [
"-y",
"@granoflow/mcp-server@latest"
],
"env": {
"GRANOFLOW_MCP_CONFIG_PATH": "/path/to/directory",
"GRANOFLOW_API_TOKEN": "<granoflow_api_token>",
"GRANOFLOW_API_BASE_URL": "https://..."
}
}
}
}{
"mcpServers": {
"granoflow": {
"command": "npx",
"args": [
"-y",
"@granoflow/mcp-server@latest"
],
"env": {
"GRANOFLOW_MCP_CONFIG_PATH": "/path/to/directory",
"GRANOFLOW_API_TOKEN": "<granoflow_api_token>",
"GRANOFLOW_API_BASE_URL": "https://..."
}
}
}
}{
"servers": {
"granoflow": {
"command": "npx",
"args": [
"-y",
"@granoflow/mcp-server@latest"
],
"env": {
"GRANOFLOW_MCP_CONFIG_PATH": "${input:granoflow_mcp_config_path}",
"GRANOFLOW_API_TOKEN": "${input:granoflow_api_token}",
"GRANOFLOW_API_BASE_URL": "${input:granoflow_api_base_url}"
}
}
},
"inputs": [
{
"id": "granoflow_mcp_config_path",
"type": "promptString",
"description": "Granoflow mcp config path",
"password": true
},
{
"id": "granoflow_api_token",
"type": "promptString",
"description": "Granoflow api token",
"password": true
},
{
"id": "granoflow_api_base_url",
"type": "promptString",
"description": "Granoflow api base url",
"password": true
}
]
}{
"mcpServers": {
"granoflow": {
"command": "npx",
"args": [
"-y",
"@granoflow/mcp-server@latest"
],
"env": {
"GRANOFLOW_MCP_CONFIG_PATH": "/path/to/directory",
"GRANOFLOW_API_TOKEN": "<granoflow_api_token>",
"GRANOFLOW_API_BASE_URL": "https://..."
}
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsoninstall npx -y @granoflow/mcp-serverversion 0.1.16
What it does
Granoflow MCP connects AI agents to a local-first task management and work memory system. It exposes tools for tracking tasks, recording reviews, and preserving lessons as memory cards, designed for agent-assisted workflows especially in software projects.
- Track what task an AI agent is working on and record its completion with meaningful reviews instead of losing context between runs.
- Preserve lessons learned from agent work as memory cards that future agent runs can search and reference to avoid repeating past mistakes.
- Review and understand agent work history so it is not a black box, allowing you to learn from or build upon previous attempts.
Tools this server exposes
| Tool | Does |
|---|---|
| granoflow_setup_write_config | Preview and configure non-default Granoflow API connection settings like port, storing non-secret values locally. |
| granoflow_agent_preferences_write_defaults | Preview or update default agent preferences for execution mode, explanation level, git workflow, and other behaviors. |
| granoflow_agent_preferences_get | Resolve a project's effective agent preferences by merging local defaults, project overrides, and newcomer-safe defaults. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.