Show HN: Run and Compare multiple autonomous browser-agents side-by-side (www.thebrowserarena.com)

A new open-source project, the-browser-arena (kalil0321/the-browser-arena), was posted on Hacker News to let you run and compare multiple autonomous browser agents side‑by‑side. It’s built around Smooth (you add a Smooth API key in Settings) and can optionally plug into major LLM providers (OpenAI, Anthropic, Google) plus a cloud browser provider (Browser‑Use) so you can run LLM-driven agents that control real browsers. Free signup credits are available for Smooth and Browser‑Use to get started quickly. This is useful for AI/ML practitioners building or evaluating agentic systems because it makes concurrent, reproducible comparisons easier: you can test different models, prompts, or tool integrations in the same environment and observe divergent behaviors in parallel. Technically, the stack relies on Smooth for orchestration, configurable LLM backends via API keys, and a browser cloud API for executing browser actions. The repo gives a hands‑on playground for debugging, benchmarking, and iterating on autonomous browsing strategies, lowering friction for experiments that require multi‑agent side‑by‑side traces and outcomes.