ZeroToken Record once. Automate forever.

✨步子哥 · 2026-03-14T01:37:38+00:00

ZeroToken - Record once. Automate forever. * { margin: 0; padding: 0; box-sizing: border-box; } body { font-family: 'Inter', -apple-system, BlinkMacSystemFont, 'Segoe UI', sans-serif; background: linear-gradient(135deg, #0a0e27 0%, #1a1f3a 50%, #0f1419 100%); color: #e8eaed; min-height: 100vh; display: flex; justify-content: center; align-items: center; padding: 20px; } .poster-container { width: 720px; min-height: 1280px; background: linear-gradient(135deg, #0d1117 0%, #161b22 100%); border-radius: 24px; overflow: hidden; box-shadow: 0 20px 60px rgba(0, 0, 0, 0.5); position: relative; } /* Background decoration */ .bg-grid { position: absolute; top: 0; left: 0; right: 0; bottom: 0; background-image: linear-gradient(rgba(88, 166, 255, 0.03) 1px, transparent 1px), linear-gradient(90deg, rgba(88, 166, 255, 0.03) 1px, transparent 1px); background-size: 40px 40px; pointer-events: none; } .bg-glow { position: absolute; top: -200px; right: -200px; width: 600px; height: 600px; background: radial-gradient(circle, rgba(88, 166, 255, 0.15) 0%, transparent 70%); pointer-events: none; } .content { position: relative; z-index: 1; padding: 48px 40px; } /* Header */ .header { margin-bottom: 48px; text-align: center; } .title { font-size: 68px; font-weight: 800; background: linear-gradient(135deg, #58a6ff 0%, #79c0ff 50%, #a5d6ff 100%); -webkit-background-clip: text; -webkit-text-fill-color: transparent; background-clip: text; letter-spacing: -2px; margin-bottom: 16px; line-height: 1.1; } .subtitle { font-size: 24px; color: #7ee787; font-weight: 600; font-family: 'Fira Code', monospace; letter-spacing: 0.5px; } /* Problem Section */ .problem-section { background: rgba(255, 107, 107, 0.08); border: 1px solid rgba(255, 107, 107, 0.2); border-radius: 16px; padding: 28px; margin-bottom: 40px; } .problem-title { font-size: 20px; font-weight: 700; color: #ff6b6b; margin-bottom: 16px; display: flex; align-items: center; gap: 10px; } .problem-content { font-size: 16px; line-height: 1.6; color: #c9d1d9; } .highlight-red { color: #ff6b6b; font-weight: 600; } /* Architecture Flow */ .architecture { margin-bottom: 40px; } .section-title { font-size: 22px; font-weight: 700; color: #58a6ff; margin-bottom: 24px; display: flex; align-items: center; gap: 12px; } .flow-container { display: flex; justify-content: space-between; align-items: center; gap: 16px; } .flow-step { flex: 1; background: rgba(88, 166, 255, 0.05); border: 2px solid rgba(88, 166, 255, 0.3); border-radius: 12px; padding: 24px 16px; text-align: center; position: relative; } .flow-step.active { background: rgba(88, 166, 255, 0.1); border-color: #58a6ff; box-shadow: 0 0 30px rgba(88, 166, 255, 0.2); } .flow-step.inactive { background: rgba(126, 231, 135, 0.05); border-color: rgba(126, 231, 135, 0.3); } .flow-icon { font-size: 36px; margin-bottom: 12px; display: block; } .flow-title { font-size: 16px; font-weight: 700; color: #e8eaed; margin-bottom: 8px; } .flow-desc { font-size: 12px; color: #8b949e; font-family: 'Fira Code', monospace; } .flow-arrow { color: #58a6ff; font-size: 28px; font-weight: bold; } /* Features Grid */ .features { display: grid; grid-template-columns: 1fr 1fr; gap: 20px; margin-bottom: 40px; } .feature-card { background: rgba(33, 38, 45, 0.6); border: 1px solid rgba(48, 54, 61, 0.8); border-radius: 16px; padding: 24px; transition: all 0.3s ease; } .feature-card:hover { border-color: rgba(88, 166, 255, 0.5); background: rgba(33, 38, 45, 0.8); } .feature-header { display: flex; align-items: center; gap: 12px; margin-bottom: 16px; } .feature-icon { font-size: 28px; color: #58a6ff; } .feature-title { font-size: 17px; font-weight: 700; color: #e8eaed; } .feature-content { font-size: 14px; line-height: 1.6; color: #8b949e; } .feature-code { background: rgba(0, 0, 0, 0.3); border-radius: 8px; padding: 10px 12px; margin-top: 12px; font-family: 'Fira Code', monospace; font-size: 11px; color: #7ee787; border-left: 3px solid #7ee787; } /* Comparison */ .comparison { background: linear-gradient(135deg, rgba(88, 166, 255, 0.05) 0%, rgba(126, 231, 135, 0.05) 100%); border: 1px solid rgba(48, 54, 61, 0.8); border-radius: 16px; padding: 28px; margin-bottom: 40px; } .comparison-grid { display: grid; grid-template-columns: 1fr 1fr; gap: 24px; } .comparison-item { padding: 20px; border-radius: 12px; } .comparison-item.old { background: rgba(255, 107, 107, 0.08); border: 1px solid rgba(255, 107, 107, 0.2); } .comparison-item.new { background: rgba(126, 231, 135, 0.08); border: 1px solid rgba(126, 231, 135, 0.2); } .comparison-title { font-size: 16px; font-weight: 700; margin-bottom: 12px; display: flex; align-items: center; gap: 8px; } .comparison-item.old .comparison-title { color: #ff6b6b; } .comparison-item.new .comparison-title { color: #7ee787; } .comparison-text { font-size: 14px; line-height: 1.5; color: #c9d1d9; } .token-waste { color: #ff6b6b; font-weight: 600; } .token-save { color: #7ee787; font-weight: 600; } /* Footer */ .footer { text-align: center; padding-top: 20px; border-top: 1px solid rgba(48, 54, 61, 0.8); } .github-box { background: rgba(88, 166, 255, 0.1); border: 1px solid rgba(88, 166, 255, 0.3); border-radius: 12px; padding: 16px 24px; display: inline-flex; align-items: center; gap: 12px; font-family: 'Fira Code', monospace; font-size: 14px; color: #58a6ff; text-decoration: none; transition: all 0.3s ease; } .github-box:hover { background: rgba(88, 166, 255, 0.15); border-color: #58a6ff; } .footer-desc { margin-top: 16px; font-size: 13px; color: #8b949e; font-style: italic; } /* Responsive */ @media (max-width: 600px) { .features { grid-template-columns: 1fr; } .comparison-grid { grid-template-columns: 1fr; } } ZeroToken Record once. Automate forever. warning 核心问题 AI Agent 控制浏览器时，每个操作都需要一次 LLM 推理。重复执行相同任务时，系统每天在重新思考同一个流程，造成巨大的 Token 浪费。 account_tree 三层架构 psychology 录制阶段 LLM 推理 → timeline 轨迹生成 Trajectory → play_circle 脚本回放 Script Engine stars 核心特性 help_outline 模糊点标记显式标记需要人工或AI判断的步骤 fuzzy_point: { requires_judgment: true } replay 确定性回放首次推理后，直接运行脚本 run_script(task_id) // 无需 LLM fingerprint 自适应定位保存元素指纹，应对网站变化 adaptive: true // 智能匹配 error_outline 统一错误契约结构化错误，便于AI决策 { retryable: true, code: "..." } compare 效率对比 close 传统方式每次执行都需 LLM 推理重复思考相同流程 Token 成本高 check_circle ZeroToken 首次推理后直接回放稳定执行确定性任务 Token 成本大幅降低 code github.com/AMOS144/zerotoken AI Agent 浏览器自动化 MCP Server

ZeroToken - Record once. Automate forever.

核心问题

AI Agent 控制浏览器时，每个操作都需要一次 LLM 推理。重复执行相同任务时，系统每天在重新思考同一个流程，造成巨大的 Token 浪费。

三层架构

录制阶段

LLM 推理

→

轨迹生成

Trajectory

→

脚本回放

Script Engine

核心特性

模糊点标记

显式标记需要人工或AI判断的步骤

fuzzy_point: { requires_judgment: true }

确定性回放

首次推理后，直接运行脚本

run_script(task_id) // 无需 LLM

自适应定位

保存元素指纹，应对网站变化

adaptive: true // 智能匹配

统一错误契约

结构化错误，便于AI决策

{ retryable: true, code: "..." }

效率对比

传统方式

每次执行都需 LLM 推理
重复思考相同流程
Token 成本高

ZeroToken

首次推理后直接回放
稳定执行确定性任务
Token 成本大幅降低

👍 1

ZeroToken Record once. Automate forever.

ZeroToken

🌟 智谱 GLM-5 已上线