Your AI has
amnesia
It sees the file you're editing but has zero awareness of the 8 files that will break when you change it.
Works seamlessly with your favorite AI tools
Numbers that matter
See how Memoria prevents production incidents before they happen
Why does your AI keep breaking things?
AI assistants are goldfish. They see the file you're editing but have zero awareness of hidden dependencies.
Thirteen engines working in parallel
Memoria pre-computes what AI cannot do efficiently—giving your assistant the intuition of a senior developer who's been on the project for years.
Volatility Engine
Scans commits for panic keywords with time-decay. Recent bugs matter more.
Entanglement Engine
Finds files that change together >15% of the time.
Sentinel Engine
Detects when coupled files are >7 days out of sync.
Static Import Engine
Uses git grep to find files that import the target.
History Search
Search git history to understand WHY code was written.
Documentation Coupling
Finds markdown files referencing your exported functions and types.
Type Coupling
Discovers files sharing type definitions via git pickaxe search.
Content Coupling
Detects files sharing string literals like error messages and constants.
Test File Coupling
Auto-discovers test and mock files matching source file naming patterns.
Environment Coupling
Finds files sharing the same environment variables (ALL_CAPS_UNDERSCORE).
Schema Coupling
Detects files affected by database schema or model changes.
API Endpoint Coupling
Finds client code calling API endpoints defined in the target file.
Transitive Coupling
Discovers files affected through barrel/index re-exports.
Total analysis time: ~150ms — all engines run in parallel
Get started in seconds
One config change. Zero API keys. Works immediately.
{
"mcpServers": {
"memoria": {
"command": "npx",
"args": ["-y", "@byronwade/memoria"]
}
}
}Or install with:
claude mcp add memorianpm i -g @byronwade/memoriaSimple Pricing
Free locally. Pay for team intelligence.
All 13 git analysis engines run free forever. Paid plans add cloud memories, guardrails, and team dashboards.
All 13 engines. No account needed.
- All 13 git analysis engines
- Unlimited local analysis
- Works offline
Cloud memories that persist.
- Unlimited cloud memories
- Personal guardrails
- Dashboard & analytics
Shared memories for your team.
- Team-wide shared memories
- Unlimited guardrails
- Org-level analytics
Frequently asked questions
Everything you need to know about Memoria
Memoria runs as an MCP (Model Context Protocol) server that your AI assistant connects to. When you edit a file, your AI calls Memoria's analyze_file tool to get risk scores, coupled files, and a pre-flight checklist. This takes ~150ms and gives your AI context it couldn't get otherwise.
The 13 git analysis engines run 100% free and locally - no account needed. Paid plans add cloud memories (shared context across your team), guardrails (file protection rules), and a dashboard. Free tier is fully functional for solo developers.
Memoria works with any AI tool that supports MCP: Claude Desktop, Claude Code, Cursor, Windsurf, Continue, Cline, and more. If your tool supports MCP servers, it supports Memoria.
The free tier runs 100% locally - your code never leaves your machine. If you opt into paid cloud features, only memories and guardrail metadata are synced, never your actual source code.
Cloud memories let your team save lessons learned (e.g., 'Safari OAuth needs a 100ms delay'). These memories appear automatically when anyone on your team edits related files, preventing the same bugs from recurring.
Guardrails are file protection rules your team sets via the dashboard. For example, block AI from editing production configs, or warn before touching critical authentication code. They enforce team policies automatically.
Still have questions? Open an issue on GitHub
Ready to give your AI memory?
Install in seconds. Zero configuration. Watch your AI stop breaking things.