Claude in Chrome
已验证Anthropic's browser agent that navigates, fills forms, and completes tasks inside Chrome
发布时间
2026
最近更新
2026-08-31
国家/地区
United States
API
不可用
自托管
否
融资
Venture-backed (Anthropic)
最近更新
2026-08
关于 Claude in Chrome
总评
The most trustworthy consumer browser agent available in 2026. It sacrifices enterprise orchestration depth for safety design and zero setup — the right call for the Claude installed base it serves.
功能特性
详细评分
优缺点
优点
- GA product from Anthropic after a year-long safety preview — not a demo
- Strongest guardrail set in the consumer browser-agent category
- Included with Claude subscriptions — no separate purchase or setup
缺点
- Needs a paid Claude subscription; no standalone pricing
- Single-browser (Chrome) scope; no team orchestration or RPA-style flows
- Sensitive-step confirmations add friction on high-volume automation
应用场景
适合谁使用?
Claude subscribers who want supervised delegation of repetitive web work inside Chrome
常见问题
What is Claude in Chrome?
It is Anthropic's browser agent: a Chrome extension in which Claude sees the page and operates the cursor and keyboard to complete multi-step web tasks under your supervision. It went generally available on August 26, 2026.
How much does Claude in Chrome cost?
There is no separate price — it requires a paid Claude subscription (Pro from $20/month), and usage draws from your plan's limits.
Is it safe to let an agent use my browser?
Anthropic treats this as the core design problem: the extension uses site-level permissions, asks for confirmation on sensitive actions, and includes defenses against prompt injection hidden in web content. Treat any browser agent as powerful but supervised software.
相关智能体
热门替代方案
与这些相似工具进行比较
链接与资源
档案由 AI 辅助、基于公开信息生成——未经独立手动测试。