minder-cli 0.6.9

Minder CLI is the command-line interface for the Minder self-hosted MCP platform.

Minder

Minder is a self-hosted MCP (Model Context Protocol) platform for repository-aware engineering intelligence.

It runs natively on macOS and Linux — no Docker, no external services. The stack is a Python FastAPI server with SQLite, Milvus Lite (embedded vector search), and llama-cpp-python for local LLM inference, distributed as a Tauri desktop app or standalone server.

What's in this repo

Component	Description
Minder Server	MCP gateway — SSE + streamable HTTP + stdio, RAG pipeline, workflow engine, memory, admin HTTP
Minder Dashboard	Astro admin console — client management, onboarding snippets, agent instructions, skill catalog, chat
Minder CLI (minder-cli on PyPI)	Edge CLI — repo sync, MCP config install, login, self-update

Architecture

Developer workstation
  ├── minder-cli          repo sync, MCP config
  └── AI agent (IDE)  ──► Minder Server :8800
                              │
                  ┌───────────┼───────────────────┐
                  │           │                   │
               SQLite    Milvus Lite        llama-cpp-python
          (users, sessions,  (embedded vector    (LLM, GGUF,
           workflows, graph)  search, file-based)  Metal/CPU)

• Transport: SSE (/sse), streamable HTTP (/mcp), stdio
• LLM inference: llama-cpp-python with GGUF models auto-downloaded from HuggingFace (Metal on Mac, CPU elsewhere)
• Vector search: Milvus Lite — embedded, file-based, no separate server
• Relational storage: SQLite (default) or PostgreSQL
• Desktop: optional Tauri v2 shell that bundles the Python server as a sidecar

Quick Start

1. Install and run

# Install dependencies and build the dashboard
make native-install

# Start the server (dashboard + MCP API on port 8800)
make native-run

GGUF models (ggml-org/gemma-4-E2B-it-GGUF) are downloaded automatically on first startup. No manual setup required.

2. Open the dashboard

http://localhost:8800/dashboard

First run → /dashboard/setup to create an admin and get your mk_... key.

3. Create a client and connect your IDE

1. Open /dashboard/clients → create a client → save the mkc_... key (shown once)
2. Open the client detail → copy the MCP snippet for your IDE from Copy-ready MCP snippets
3. Open /dashboard/instruction → copy the agent orchestration rules for your IDE

4. Install the CLI and sync a repository

# Install
uv tool install minder-cli

# Log in
minder login --client-key mkc_your_key --server-url http://localhost:8800/sse

# Write MCP config to your IDE (optional — dashboard shows it too)
minder install --target vscode --target claude-code

# Sync a repository
minder sync

MCP Tools

When connected, Minder exposes these tools to your AI agents:

Tool	Description
minder_query	Full RAG pipeline: retrieve → reason → verify → respond
minder_search_code	Semantic code search across indexed repos
minder_search_errors	Look up past error patterns
minder_find_impact	Find what a change might affect
minder_memory_store / minder_memory_recall	Persistent engineering memory
minder_session_create / minder_session_save / minder_session_restore	Cross-machine session continuity
minder_workflow_get / minder_workflow_step / minder_workflow_guard	Workflow governance
minder_skill_store / minder_skill_recall	Reusable pattern catalog
minder_agent_list / minder_agent_get	SubAgent registry

Dashboard Pages

Route	Description
/dashboard	Home — stats and quick nav
/dashboard/clients	Create clients, copy MCP snippets
/dashboard/instruction	Agent orchestration rules — copy for Claude Code, Cursor, VS Code
/dashboard/sessions	LLM session management
/dashboard/memories	Persistent memory browser
/dashboard/skills	Skill / pattern catalog
/dashboard/agents	SubAgent registry
/dashboard/chat	Browser-based runtime chat
/dashboard/repositories	Repo graph explorer
/dashboard/workflows	Workflow definitions
/dashboard/observability	Audit and trace

Documentation

• Local Dev Setup
• Production Deployment
• System Design
• Admin & Client Onboarding
• Minder CLI

License

Apache License 2.0. See LICENSE for details.