Alternatives to WebsiteBench

Open-source Python library that makes websites accessible to AI agents, letting them drive a real browser to click, type, and fill forms.

Open-source browser API for agents. Cloud Launch is $0 plus usage with a one-time $30 credit. Scale is $250/month plus usage.

Cloudflare agent browser on Workers. GitHub 404. Not an open-source Chromium clone.

Open-source Microsoft sandbox and benchmark that verifies AI agents reliably complete stateful business tasks, exposing the discovery-reliability gap.

Hosted browser infra for agents. Developer is $20/month. Startup is $99/month. Scale starts at $249/month after a custom form.

Tencent's MIT-licensed CLI plus browser extension lets a coding agent borrow one of your already signed-in tabs, then hand it back when the task is done.

Open-source LLM tracing and evals. Cloud Hobby is $0. Core is $29/seat. Pro is $199/seat. Plus usage after included units.

Open-source Rust framework and optional durable platform for building AI agents with typed tools, persistent sessions, and controlled production channels.

Open-source capability layer that gives AI agents internet access to Twitter, Reddit, YouTube, Bilibili, and XiaoHongShu from one CLI, with local cookies.

Open-source long-term memory for coding agents that preserves project context across agent CLIs, machines, and team handoffs.

Buzz is Block's open-source Nostr workspace where humans and AI agents share the same channel, with every message signed and every agent scoped by its own key.

An open-source CLI and optional MCP server that gives AI agents cited, ranked answers from large local record sets.

Open-source workspace that assigns issues to coding agents as teammates. Self-hostable, 26 CLIs, daemon runs on your machine.

Open-source sandbox service for agent-authored code that reports its actual isolation tier, enforces a minimum tier, and keeps API credentials outside the run.