该浏览器智能体将可见页面控件整理成带索引的操作空间,用于执行自然语言目标。 Jev 在一次请求中选择操作及兼容元素;只有 TYPE_TEXT 的文本内容由独立的小型大语言模型生成。
浏览器自动化
用类型化决策驱动 Chrome 或 Playwright
287 个项目 · 全部项目 →
Hermes Jev Skills 提供模型路由、搜索、记忆筛选、消息分流和浏览器操作等可复用技能。 Jev 负责选择模型、排序检索片段等有界决策;宿主模型负责生成文字。 技能适用于 Hermes、Claude Code 和 Codex;影子模式只记录路由决策,不切换模型。
DSCODE 是 macOS 编程智能体,终端界面、命令行和脚本共享持久会话运行时。 仓库包含独立的 Jev 客户端,连接 OpenRouter alpha Decisions 端点,支持 Choice、Score 和 Noul。 可在会话中编辑、测试和向其他智能体交接任务;Jev 客户端需要 API 密钥。
Jev Browser Use 接管导航、点击、开关和滚动,文字输入与最终验证由 Codex 负责。 Jev 根据无障碍文本从已观察控件中选择动作,由现有 Computer Use 连接执行。 需要 Node.js 22+、已连接浏览器的 Codex,以及匹配的 TypeSafe 或 OpenRouter 凭据。
Surf 将 Chrome 导航、页面读取、点击和截图开放为智能体及脚本可用的命令行操作。 可选语义命令让 Jev 查找控件、排序页面区域、验证结果并选择有界动作。 已知流程可用确定性命令;semantic.act 按目标探索,并检查置信度、元素时效与预算。
jev-voice-browser 将语音转录流发送给 Node 服务,通过 Playwright 控制可见 Chromium 窗口。 Jev 选择意图和目标;代码提取原文候选,并决定执行、等待还是询问。 需要 Node 20+、TypeSafe 密钥,以及 Chrome 或 Edge 麦克风控制页。
Unclutter 在 Chrome、Chromium 和 Firefox 中隐藏非必要元素,并为匹配的页面模板保存规则。 Jev 将候选元素分为类型化类别;扩展代码校验响应、控制选择器,并以可撤销方式隐藏内容。 选择 TypeSafe 或 Vercel 凭据,默认手动分析;已保存规则可继续离线使用。
仓库提供面向 Jev 风格 Choice、Score 和 Noul 工作流的开放模型,区分 Verdict 与 Verdict 2.0。 用 Python HTTP 服务器托管 webgpu-demo,打开文档中的浏览器演示;模型产物与评估记录另有链接。
Jev Browser 从起始 URL 执行任务,返回最终页面、截图和决策轨迹。 准备 Node.js 22 和 Jev 服务密钥,使用 npm CLI、MCP 服务器或 navigate 库。
A Browser Use Agent that actually reads social media. Fast. Precise. Deep.
System 1 decision models (Jev, Laya, Cua-S1) as brain for agents: Browser use, computer use, games and robotics
AI-assisted Windows and Chrome GUI delegation for Codex: task contracts, local execution, Jev semantic decisions, recovery and outcome verification.
Open-source, local-first social media research agent for Instagram, TikTok, and LinkedIn. Jev routes read-only steps; socai CLI captures cited browser evidence.
Non-autoregressive decision engine on ModernBERT (151M) with calibrated uncertainty (RLCD), TypeSafe AI Jev benchmark audit, and in-browser WebGPU playground
A fast browser agent: Jev picks each action from what is on the page, an LLM reads and plans, and every claim in an answer cites a quote from the page.
Jev Browser — indexed browser automation. Jev chooses the control, Playwright acts. A CodexQA skill.
Browser automation where an LLM plans and Jev (Typesafe System One) decides. Library, CLI and MCP server.
🧹 Fun project: a Chrome extension that asks a tiny AI decision model (TypeSafe Jev) "is this DOM element an ad?" and pops it off the page. BYOK, no backend, not a real ad blocker.
Everything you need to run TypeSafe's Jev with Claude Code: a tool-call guard, tier guard, file search, browser agent, review, belay, compaction and installers.
Chrome extension: vibe-check your X posts with TypeSafe's Jev before you hit Post
Practical, tested recipes for TypeSafe's Jev decision model on OpenRouter: support triage, database indexing, file organizing, tagging, taxonomies, dedupe, PII detection, extraction, search re-ranking and a browser agent.
Data extraction for systematic reviews, quoted from the papers. Ask a trial report and its supplements your extraction form or a RoB 2, ROBINS-I, QUADAS-2 or TIDieR template; Jev points at the lines, every answer is a verbatim quote with its page, you check it and export the table. Files stay in your browser.
Jev for Chrome: drives the tab you are looking at with TypeSafe Jev, a sub-second decision model. Community port of browser-use/jev-ultrafast, not affiliated with TypeSafe.
A browser-native Doom agent experiment with structured spatial state, composable AI controls, live decision telemetry, and a Chocolate Doom WebAssembly runtime.