Installation

Simple Install

The easiest way to get started is with the reactive-agents meta-package, which bundles everything:

# Bun (recommended)
bun add reactive-agents effect

# npm
npm install reactive-agents effect

# pnpm
pnpm add reactive-agents effect

Then import from a single entry point:

import { ReactiveAgents } from "reactive-agents";

Modular Install

The framework is modular — install only the packages you need:

Package	Description	Required?
@reactive-agents/core	EventBus, AgentService, TaskService, types	Yes
@reactive-agents/runtime	ExecutionEngine, ReactiveAgentBuilder	Yes
@reactive-agents/llm-provider	LLM adapters (Anthropic, OpenAI, Gemini, Ollama, LiteLLM 40+)	Yes
effect	Effect-TS runtime	Yes
@reactive-agents/memory	Working, Semantic, Episodic, Procedural memory	Optional
@reactive-agents/reasoning	ReAct, Plan-Execute, Tree-of-Thought, Reflexion	Optional
@reactive-agents/tools	Tool registry, sandbox, MCP client	Optional
@reactive-agents/guardrails	Injection, PII, toxicity detection	Optional
@reactive-agents/verification	Semantic entropy, fact decomposition	Optional
@reactive-agents/cost	Complexity routing, budget enforcement	Optional
@reactive-agents/identity	Agent certificates, RBAC	Optional
@reactive-agents/observability	Tracing, metrics, structured logging	Optional
@reactive-agents/interaction	5 interaction modes, checkpoints	Optional
@reactive-agents/orchestration	Multi-agent workflows	Optional
@reactive-agents/a2a	Agent-to-Agent protocol, Agent Cards, JSON-RPC, SSE streaming	Optional
@reactive-agents/prompts	Template engine, built-in prompt library	Optional

bun add @reactive-agents/core @reactive-agents/runtime @reactive-agents/llm-provider effect

Environment Variables

Create a .env file:

# LLM Provider — set at least one
ANTHROPIC_API_KEY=sk-ant-...        # Anthropic Claude
OPENAI_API_KEY=sk-...               # OpenAI GPT-4o
GOOGLE_API_KEY=...                  # Google Gemini

# Tools (optional)
TAVILY_API_KEY=tvly-...             # Enables built-in web search tool

# Embeddings (for Tier 2 semantic memory)
EMBEDDING_PROVIDER=openai           # "openai" | "ollama"
EMBEDDING_MODEL=text-embedding-3-small

# Tuning (optional)
LLM_DEFAULT_MODEL=claude-sonnet-4-20250514
LLM_DEFAULT_TEMPERATURE=0.7
LLM_MAX_RETRIES=3
LLM_TIMEOUT_MS=30000

TypeScript Configuration

Reactive Agents requires TypeScript 5.5+ with strict mode:

{
  "compilerOptions": {
    "target": "ES2022",
    "module": "ESNext",
    "moduleResolution": "bundler",
    "strict": true,
    "exactOptionalPropertyTypes": true,
    "noUncheckedIndexedAccess": true
  }
}