MaestroError/LarAgent

What it solves

LarAgent provides a developer-friendly framework for integrating AI agents into Laravel applications. It eliminates the boilerplate required to build, maintain, and scale AI agents by providing a structured, Eloquent-like API that integrates seamlessly with the Laravel ecosystem.

How it works

Developers create agents by extending a base Agent class and defining properties for the LLM model, provider, and instructions. The framework handles the heavy lifting of conversation history, tool integration, and provider management. It uses an attribute-based system for defining tools (functions the agent can call) and supports multiple storage options for chat history (such as cache or JSON).

Who it’s for

Laravel developers who want to build AI-powered features, conversational interfaces, or internal automation tools without leaving the familiar Laravel design patterns.

Highlights

• Eloquent-style API: Create agents using a syntax familiar to Laravel developers, including artisan commands for generation.
• Multi-provider Support: Built-in support for multiple LLM providers with automatic failover capabilities.
• Tooling System: Easy tool creation via class methods or dedicated classes, including support for MCP (Model Context Protocol) servers.
• Flexible Memory: Pluggable chat history management with per-user tracking and multiple storage drivers.
• Structured Output: Support for reliable, structured data responses and multi-modal input (images).
• Observability: An extensive event system that allows developers to hook into nearly every agent interaction.