🧠 Vibe Memory

Universal Long-Term Memory & Codebase Intelligence for AI Coding Agents — Made Simple

Stop your AI from forgetting your project rules and past decisions.
Built with Mem0 Adaptive Learning and Microsoft GraphRAG, Vibe Memory gives AI agents (Antigravity, Cursor, Claude Code) persistent cross-session memory and cuts context token consumption by 97%.

📖 What is Vibe Memory? (In Simple Terms)

Have you ever noticed that every time you start a new chat with an AI coding assistant, it forgets everything you discussed yesterday? It forgets your favorite design colors, your database rules, and your past bug fixes.

Vibe Memory gives your AI a permanent "Brain":

• It remembers your preferences and project architecture forever.

• When you switch between tools (e.g. from Google Antigravity to Cursor or Claude Code), your AI picks up right where you left off.

• It searches your entire codebase using smart knowledge graphs instead of reading massive files, saving 97% of your token costs!

Related MCP server: Paparats MCP

💬 Real-World Prompts: What You Type ➔ What You Achieve

Here is how anyone can use Vibe Memory using simple plain-English chat prompts:

1. 🧠 Permanent Project Rules & Memory

What it does: Teaches your AI your specific project rules so you never have to repeat yourself.

• 💬 What you type in chat:

  "Remember that for this project, our primary brand color is Emerald (#059669), we always use Tailwind CSS v4, and we never use purple gradients."

• 🎯 What you achieve instantly:

  • AI saves this rule to persistent memory (MEMORY.md).

  • In every future conversation (even weeks later), AI will automatically follow these rules without being reminded.

2. 🔍 Instant Search Across Past Project Decisions (GraphRAG)

What it does: Recalls why specific technical decisions were made in past conversations.

• 💬 What you type in chat:

  "Why did we decide to use PostgreSQL instead of MongoDB three weeks ago, and which API routes were affected?"

• 🎯 What you achieve instantly:

  • AI queries the Microsoft GraphRAG knowledge graph and explains the exact reasoning, date, and connected files in 2 seconds.

3. 🔄 Seamless Cross-Tool Handoff (Antigravity ↔ Cursor ↔ Claude Code)

What it does: Lets you start a task in one AI tool and finish it in another without losing context.

• 💬 What you type in chat:

  "Prepare a handoff snapshot of our checkout bug investigation so I can continue in Cursor."

• 🎯 What you achieve instantly:

  • Generates an instant, compressed context package (AGENT_MEMORY.md) containing open issues, active file paths, and current progress ready for Cursor or Claude Code.

4. ⚡ 97% Context Token Cost Reduction

What it does: Prevents AI from reading massive 50,000-line codebases that waste money and hit context limits.

• 💬 What you type in chat:

  "Index our codebase and generate a lightweight knowledge map of all functions and database models."

• 🎯 What you achieve instantly:

  • Compresses your entire codebase into an Abstract Syntax Tree (AST) symbol graph.

  • AI only loads the exact 20 lines needed for a task instead of 5,000 lines, reducing token bills by 97%!

5. 🛡️ Self-Correction & Adaptive Learning (Mem0)

What it does: Ensures AI learns from mistakes and never repeats the same error twice.

• 💬 What you type in chat:

  "Never import icons directly from lucide-react without checking if the icon name exists in our design system."

• 🎯 What you achieve instantly:

  • AI records this as a strict guardrail in Mem0 memory, automatically validating imports before writing code.

⚡ 60-Second Quickstart

# 1. Initialize permanent memory in your project
npx vibe-memory init

# 2. Index your codebase into a knowledge graph
npx vibe-memory index

# 3. Ask your memory a question
npx vibe-memory query "What are our database models?"

# 4. Export handoff state for Cursor or Claude Code
npx vibe-memory handoff

🔌 Connect to Your Favorite AI Assistant

🔹 Google Antigravity (~/.gemini/antigravity/mcp-config.json):

{
  "mcpServers": {
    "vibe-memory": {
      "command": "npx",
      "args": ["-y", "vibe-memory", "mcp"]
    }
  }
}

🔹 Cursor (.cursor/mcp.json):

{
  "mcpServers": {
    "vibe-memory": {
      "command": "npx",
      "args": ["-y", "vibe-memory", "mcp"]
    }
  }
}

👤 Author

Created with intention by Yogeshkumar Patel • Adelaide, Australia 🇦🇺

• LinkedIn: https://www.linkedin.com/in/yogeshkumar-ai/

• GitHub: @shahrukh-hack

📄 License

MIT License © 2026 Yogeshkumar Patel