> ## Documentation Index
> Fetch the complete documentation index at: https://icrl.dev/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# TypeScript Package API

> Overview of the ICRL TypeScript library

## Overview

The ICRL TypeScript package (`icrl`) provides the same trajectory-learning framework as the Python package. Use the pages in this section for detailed API documentation.

## Core Components

* [Agent](/docs/api-reference/typescript-agent) — main agent class
* [TrajectoryDatabase](/docs/api-reference/typescript-trajectory-database) — trajectory store with semantic search
* [TrajectoryRetriever](/docs/api-reference/typescript-trajectory-retriever) — step-level retrieval
* [CurationManager](/docs/api-reference/typescript-curation-manager) — automatic curation
* [ReActLoop](/docs/api-reference/typescript-react-loop) — planning, reasoning, acting loop

## Storage & Protocols

* [StorageAdapter](/docs/api-reference/typescript-storage) — pluggable storage interface; built-in `FileSystemAdapter`
* [Environment](/docs/api-reference/typescript-environment) — environment protocol
* [LLMProvider](/docs/api-reference/typescript-llm-provider) — LLM protocol
* [Embedder](/docs/api-reference/typescript-embedder) — embedding protocol

## Models

* [Models](/docs/api-reference/typescript-models) — `Trajectory`, `Step`, `Message`, `StepExample`, helpers

## Providers

* [OpenAI Provider](/docs/api-reference/typescript-openai-provider) — `OpenAIProvider`, `OpenAIEmbedder`
* [Anthropic Provider](/docs/api-reference/typescript-anthropic-provider) — `AnthropicProvider`, `AnthropicVertexProvider`