LarAgent v0.4 is here — Gemini, Streaming, Fallbacks & More! 🚀
We’re excited to announce the release of LarAgent v0.4.0, bringing major upgrades that make your Laravel-based AI agents faster, smarter, and more resilient.
Let’s break down what’s new 👇
🔹 Gemini Driver Support
Now you can use Google’s Gemini API alongside OpenAI with zero hassle. Swap drivers effortlessly and explore multiple LLMs in your projects.
use LarAgent\Drivers\OpenAi\GeminiDriver;🔹 Streaming Responses
Get real-time output as the model generates it. Perfect for chatbots or anything interactive. No more waiting for full responses.
$stream = YourAgent::forUser($user)->respondStreamed($input);🔹 Fallback Provider Logic
Your agent now auto-switches to a fallback provider when the primary fails — due to rate limits, downtime, or network errors.
Set it in your config:
'fallback_provider' => 'default'🔹 Flexible Configuration Per Agent
Need different settings for each agent? Now you can override apiKey, apiUrl, or fetch them dynamically from your DB or tenant logic.
public function getApiKey(): string
{
return auth()->user()->apiKey;
}🔹 Tooling Upgrades
You can now add/remove tools dynamically at runtime using either class references or objects — no need to hardcode.
$agent->withTool(MyTool::class);
$agent->removeTool(MyTool::class);This release is packed with enhancements to make your agents more modular, reliable, and developer-friendly.
Check the full docs here 👉 https://docs.laragent.ai
Subscribe to learn more about AI development with Laravel 💪
Happy building,
The LarAgent Team 💥