Introduction

AgenticFORGE is a TypeScript framework for building production-ready AI agents. It is centered around tool invocation as the primary execution primitive, and ships with:

• 5 classic agent workflow implementations (ReAct, Plan-and-Solve, Reflection, FunctionCall, Simple)
• A composable multi-type memory system (working, episodic, semantic, perceptual)
• A built-in RAG pipeline with pluggable vector storage
• A token-aware context builder for LLM input window management
• A set of ready-to-use built-in tools (search, memory, notes, RAG, terminal)

Why AgenticFORGE?

Most agent frameworks are either too opinionated (locked to one LLM provider or one architecture) or too low-level (you wire everything yourself). AgenticFORGE sits in the middle:

• Opinionated where it matters: clean agent loop abstractions, typed tool contracts, structured memory
• Flexible where it counts: swap LLM providers, storage backends, or memory types without rewriting logic
• TypeScript-first: full type safety from tool parameters to agent outputs

Package Architecture

AgenticFORGE is a monorepo of focused packages:

Package	Size	Role
@agenticforge/kit	1.8 KB	All-in-one entry point
@agenticforge/core	18 KB	LLM client, base types
@agenticforge/agents	14.5 KB	5 agent implementations
@agenticforge/memory	multi-entry	Memory manager + RAG
@agenticforge/tools	4.3 KB	Tool abstraction layer
@agenticforge/tools-builtin	45 KB	5 ready-to-use tools
@agenticforge/context	2.5 KB	Context builder
@agenticforge/utils	0.9 KB	Utilities

Next Steps

• Quick Start — up and running in 60 seconds
• Installation — install options and sub-path imports
• Agents Guide — choose the right agent workflow