loomgit 0.1.6

AI-powered context engine that weaves developer memory and codebase history

loomgit

The AI-Powered Developer Memory & Context Engine

loomgit automatically captures, structures, and semantically indexes your coding decisions, bug fixes, architecture choices, and lessons learned into an AI-queryable memory graph.

📋 Prerequisites

Before installing loomgit, make sure you have:

• Python 3.11+ installed
• Git installed and initialized in your project
• Groq API Key (Free at console.groq.com)
• Google Gemini API Key (Free at aistudio.google.com)

---

🚀 Step-by-Step Getting Started Guide

Follow these sequential steps to set up loomgit, configure keys, capture memories, launch the web UI, and connect AI coding assistants.

1️⃣ Installation

pip install loomgit

Or install from source for development:

git clone https://github.com/Mohamedarsath26/loomgit.git
cd loomgit
pip install -e .

---

2️⃣ Configure API Keys

Set up your Groq (Llama 3.3 70B) and Google Gemini (Embeddings) keys:

loomgit setup

---

3️⃣ Enable Auto-Git Capture (Git Hook)

Install the automatic post-commit hook in your repository:

# Run inside your project repository:
loomgit install-hook

How it works: Once installed, loomgit automatically triggers on every commit, whether made via:

• 💻 Terminal / CLI: Manual git commit -m "..."
• 🎨 VS Code / IDE GUI: Commit button in Source Control UI, Cursor, or JetBrains IDEs

---

4️⃣ Connect AI Coding Assistants (MCP Setup)

Connect loomgit to your AI pair programmer:

🌌 Option A: Connect to Google DeepMind Antigravity IDE

loomgit setup-antigravity

🤖 Option B: Connect to Claude Code / Claude Desktop

loomgit setup-claude

---

5️⃣ Launch Local Web Dashboard

Explore your memory timeline, search past decisions, and filter by project scope:

loomgit ui

🧠 Starting loomgit Local Web Dashboard...
✓ Server running at http://127.0.0.1:8001

View Live Dashboard: Open http://127.0.0.1:8001 in your web browser.

---

💻 CLI Command Reference

Command	Description	Example
loomgit log	Log a manual memory entry	loomgit log "Switched to gRPC for speed"
loomgit search	Semantic + Keyword search	loomgit search "gRPC migration"
loomgit list	View chronological timeline	loomgit list --limit 10
loomgit ui	Launch local web dashboard	loomgit ui
loomgit install-hook	Enable Git post-commit hook	loomgit install-hook
loomgit setup-antigravity	Auto-connect MCP to Antigravity IDE	loomgit setup-antigravity
loomgit setup-claude	Auto-connect MCP to Claude Code	loomgit setup-claude

---

⚖️ Benchmark & Feature Comparison

loomgit is designed specifically for developer memory and codebase intent tracking. Below is a side-by-side comparison against direct open-source alternatives:

Feature / Capability	loomgit	aicommits / OpenCommit	Supermemory / Mem0	Timescale Memory Engine	Graphify
Primary Focus	Developer Memory & Intent	Commit message generator	Generic LLM memory	PostgreSQL agent memory	Codebase AST graph
Automatic Git Hook Capture	⚡ Yes (loomgit install-hook)	⚠️ Manual git commit run	❌ No	❌ No	❌ No
AI Intent & Reasoning Analysis	🧠 Llama 3.3 70B	❌ Short message only	⚠️ Generic facts	⚠️ Unstructured	❌ Code structure only
Semantic Vector Search Engine	🔍 Gemini + Qdrant Vector DB	❌ No search	⚠️ Vector only	⚠️ pgvector only	⚠️ Graph traversal
Smart Per-File Diff Budgeting	🎯 Yes (800-char/file)	❌ Truncate blob	❌ No diff parsing	❌ No diff parsing	❌ No diff parsing
1-Click MCP Setup (Antigravity/Claude)	🔌 Yes (setup-antigravity)	❌ No	⚠️ Partial API	🔌 Yes	⚠️ Partial
Standalone Local Web Dashboard	📊 Yes (loomgit ui)	❌ No	❌ SaaS / Cloud	❌ No	⚠️ Static HTML report
Local-First Zero External DB	💻 SQLite + Local Qdrant	N/A	❌ Cloud / External	❌ Requires Postgres	💻 Local AST

---

🏛️ Architecture & How It Works

                     ┌──────────────────────────────────────────┐
                     │          Developer Activity              │
                     │  (Git Commit / CLI `loomgit log`)       │
                     └────────────────────┬─────────────────────┘
                                          │
                                          ▼
                     ┌──────────────────────────────────────────┐
                     │          Extraction Pipeline             │
                     │  • Per-file Smart Diff Budgeter          │
                     │  • Groq Llama 3.3 70B Engine             │
                     └────────────────────┬─────────────────────┘
                                          │
                        ┌─────────────────┴─────────────────┐
                        ▼                                   ▼
          ┌──────────────────────────┐        ┌──────────────────────────┐
          │  SQLite Store            │        │  Qdrant Vector DB        │
          │  • Chronological Event   │        │  • Google Gemini 768d    │
          │  • Metadata & History    │        │    Dense Embeddings      │
          └─────────────┬────────────┘        └─────────────┬────────────┘
                        │                                   │
                        └─────────────────┬─────────────────┘
                                          │
                                          ▼
                     ┌──────────────────────────────────────────┐
                     │          Query & Access Interfaces       │
                     │  • Local Web UI (`loomgit ui`)          │
                     │  • MCP Server (Antigravity / Claude)    │
                     │  • Terminal CLI (`loomgit search`)       │
                     └──────────────────────────────────────────┘

---

🗺️ Future Roadmap & Planned Features

loomgit is actively evolving! Planned upcoming capabilities include:

• 🤖 Multi-LLM Provider Engine: Support for interchangeable extraction models beyond Groq:
  • OpenAI (gpt-4o, gpt-4o-mini)
  • Anthropic Claude (claude-3-5-sonnet)
  • DeepSeek (deepseek-chat, deepseek-r1)
  • Local LLMs via Ollama / vLLM (100% offline extraction)
• 🔌 Expanded Agentic CLI & MCP Integrations:
  • OpenCode / Cursor / Windsurf / Aider / Roo Code MCP auto-discovery
  • Automated memory recall during pull request reviews
• 🕸️ Graph-Based Relationship Memory: Temporal knowledge graphs mapping file-to-decision dependencies over time.

---

📄 License

Distributed under the MIT License. See LICENSE for details.