MCP Catalogs
Home

Wax

by christopherkarani·734·Score 53

Swift-native memory layer for AI agents providing high-performance RAG on Apple Silicon with zero configuration.

ai-llmdeveloper-toolsdatabase
44
Forks
2
Open issues
this month
Last commit
2d ago
Indexed

Overview

Wax is a high-performance, single-file memory layer for AI agents on Apple platforms. It stores documents, embeddings, and structured knowledge in a single portable .wax file. Designed for sub-millisecond recall latency with hybrid search capabilities (text + vector), it operates entirely on-device without server or cloud dependencies. The implementation leverages Metal acceleration for vector search and SQLite FTS5 for text search, making it particularly optimized for Apple Silicon hardware.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

you:Persistent memory for AI coding assistants like Claude Code and Cursor
you:Local RAG implementation for private on-device AI applications
you:Session memory management across multiple AI assistant interactions
you:What platforms does Wax support?
you:How does Wax ensure data privacy?

When to choose this

Choose Wax when building Apple-specific AI agents that need fast, private memory without cloud dependencies or complex setup.

When NOT to choose this

Don't choose Wax if you need cross-platform support, cloud-based deployment, or have non-Apple hardware requirements.

Tools this server exposes

12 tools extracted from the README
  • handoff_latest

    Load prior context from previous sessions

  • session_start

    Start a new memory session

  • remember

    Store memories, facts, and decisions

  • recall

    Retrieve assembled context from memory

  • search

    Perform hybrid or text-only search in memory

  • handoff

    Pass session content with optional project and pending tasks

  • session_end

    End the current memory session

  • corpus_search

    Search across broker-managed session history

  • entity_upsert

    Insert or update structured entities

  • fact_assert

    Assert a structured fact

  • fact_retract

    Remove a structured fact

  • facts_query

    Query structured facts

Comparable tools

sqlite-mcpvector-mcplance-mcpchromadb-mcp

Installation

Installation

Swift Package Manager

dependencies: [
    .package(url: "https://github.com/christopherkarani/Wax.git", from: "0.1.8")
]

MCP Server Installation (Claude Code)

npx -y waxmcp@latest mcp install --scope user

Claude Desktop Configuration

Add the following to your Claude Desktop configuration file:

{
  "mcpServers": {
    "wax": {
      "command": "wax-mcp",
      "args": []
    }
  }
}

FAQ

What platforms does Wax support?
Wax supports iOS and macOS, optimized for Apple Silicon (M-series) processors.
How does Wax ensure data privacy?
Wax operates entirely on-device with no server or cloud dependencies, ensuring 100% local data privacy.

On Hacker News

Recent discussion from the developer community.

Compare Wax with

GitHub →

Last updated · Auto-generated from public README + GitHub signals.