memor-cli
v0.1.11
Published
Software is alive. See it. Interactive architecture projection for any TS/JS codebase.
Downloads
783
Maintainers
Readme
Memor
Instant codebase understanding for any TS/JS project.
Memor analyzes your repository and generates an interactive architectural briefing — structure, runtime flows, and change impact — all in your browser, in seconds.
No cloud. No AI API calls. Everything runs locally on your machine.
Quick Start
npx memorThat's it. Memor analyzes your codebase and opens an interactive app at http://localhost:4173.
Install (optional)
npm install -g memor-cliThen run memor from any TS/JS project directory.
Options
memor # Analyze current directory
memor /path/to/project # Analyze a specific project
memor --port 3000 # Use a custom port
memor --open # Auto-open browser after analysisWhat You Get
Overview
A snapshot of what the project is — identity, tech stack, architecture zones, entry point, and a flow preview. One card, instant context.
Structure
An interactive radial map of the codebase architecture. Navigate zones, drill into systems, see connections and relationships. Progressive disclosure — zones first, then systems inside each zone.
Flow
Animated storytelling of how the system actually works at runtime. Detects architectural patterns (request pipelines, data flows, dev loops) and presents them as step-by-step timelines with system context, file paths, and fused impact intelligence.
Impact
"If I change THIS, what breaks?" Select any system and see a directional ripple graph showing direct and indirect downstream impact, risk levels, blast radius, and affected areas.
How It Works
Memor scans your project structure and source files to detect:
- Systems — logical modules, packages, or architectural boundaries
- Zones — groups of related systems (e.g., "Core Runtime", "API Layer", "Testing")
- Connections — import relationships, extensions, and coupling between systems
- Flows — runtime behavior patterns based on detected archetypes
- Impact — change propagation paths using the connection graph
Everything is deterministic. Same codebase, same result.
MCP Integration
Memor can run as an MCP (Model Context Protocol) server, letting your AI assistant query the architecture directly from your editor.
Setup
Add to your Cursor/VS Code MCP config:
{
"memor": {
"command": "memor-mcp",
"args": ["/path/to/your/project"]
}
}Available Tools
| Tool | Description |
|------|-------------|
| get_architecture | High-level architecture overview — repo mode, zones, couplings |
| list_systems | All detected systems with type, tier, and description |
| get_system_detail | Deep dive into a specific system — connections, blocks, internal structure |
| get_impact | Change impact analysis — blast radius, risk levels, affected systems |
| get_flows | Runtime flows — how the system actually works step by step |
| get_zones | Architectural zones and their member systems |
| search_systems | Keyword search across system names, types, tech stacks |
| get_reading_order | Recommended order for reading the codebase |
| get_couplings | System coupling relationships ranked by strength |
Available Prompts
| Prompt | Description |
|--------|-------------|
| understand-codebase | Architecture-aware understanding prompt |
| plan-change | Impact-aware change planning for a specific system |
Requirements
- Node.js >= 18
- Works on any TS/JS codebase (monorepos, frameworks, libraries, web apps, APIs)
Contributing
Contributions are welcome! Please open an issue first to discuss what you'd like to change.
