Pinned Loading

1. Wax Wax Public

   Single-file memory layer for AI agents, sub mili-second RAG on Apple Silicon. Metal Optimized On-Device. No Server. No API. One File. Pure Swift

   Swift 687 30

2. Swarm Swarm Public

   🐦‍🔥 A production ready, enterprise-grade Agent Orchestration Framework written in pure Swift 6.2

   Swift 407 25

3. Espresso Espresso Public

   Train and run transformers directly on Apple's Neural Engine in Swift

   Swift 89 4

4. Conduit Conduit Public

   🦑 Unified Swift SDK for LLM inference across local and cloud providers

   Swift 80 6

5. Hive Hive Public

   Deterministic graph runtime for agent workflows in Swift. Same input, same output, every time — golden-testable, checkpoint-resumable, and built entirely on Swift concurrency.

   Swift 19 4

6. EdgeRunner EdgeRunner Public

   🚀 LLM inference Engine in Swift/Metal, Load GGUF and safe tensors modes, no conversion, no cpp, pure swift

   Swift 20