Intelligence, Forged.
Ferrum Labs crafts metal-strong building blocks for AI in Rust. Our ecosystem powers composable, reliable reasoning systems designed for performance and clarity.
FerricLink
LangChain-inspired, Rust-native framework for composing LLM tools, retrievers, and reasoning steps into reliable pipelines.
- 🧩 Modular LinkNodes with strong typing
- 📦 Ergonomic error handling and observability hooks
- 🔌 Plug-ins for providers & vector stores
FerricMap
LangGraph-inspired orchestration engine that treats reasoning as a directed graph with controllable state, memory, and tools.
- 🕸️ Deterministic node/edge traversal
- 🔁 Cycles, branches, and checkpoints
- ⚡ Async-friendly execution core
FerricMem
Associative memory and retrieval system that combines vector and structured memory adapters for intelligent recall in Rust.
- 🧠 Memory persistence and context retrieval
- 📚 Integrates seamlessly with FerricLink
- 🔍 Vector and symbolic hybrid search
Forged, not glued.
We believe AI systems should be composed from clear, interoperable parts — chains, maps, and memory — each engineered with Rust’s safety guarantees. The result is software that scales in capability and reliability.