经 AI Skill Hub 精选评估,智能AI工作流 获评「强烈推荐」。这款Agent工作流在功能完整性、社区活跃度和易用性方面表现出色,AI 评分 8.0 分,适合有一定技术背景的用户使用。
智能AI工作流 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
智能AI工作流 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
# 方式一:cargo install(推荐) cargo install senweavercoding # 方式二:从源码编译 git clone https://github.com/senweaver/SenWeaverCoding cd SenWeaverCoding cargo build --release # 二进制在 ./target/release/senweavercoding
# 查看帮助 senweavercoding --help # 基本运行 senweavercoding [options] <input> # 详细使用说明请查阅文档 # https://github.com/senweaver/SenWeaverCoding
# senweavercoding 配置说明 # 查看配置选项 senweavercoding --config-example > config.yml # 常见配置项 # output_dir: ./output # log_level: info # workers: 4 # 环境变量(覆盖配置文件) export SENWEAVERCODING_CONFIG="/path/to/config.yml"
AI Code Editor & Autonomous Agent Runtime · Rust back-end + Tauri/React desktop
SenWeaverCoding is a desktop AI code editor that ships the entire agent runtime as a single installable application. The back-end is a Rust library (src/) embedded directly into a Tauri 2 shell; the front-end is a React + Vite UI living in desktop/. There is no sidecar process, no remote service to talk to, and no separate CLI install — one installer gives you the full IDE and a terminal-callable sen command.
---
| Area | What you get |
|---|---|
| **Desktop-first** | Tauri 2 shell with native menus, multi-tab sessions, in-app PTY terminal, file browser, and an embedded browser dock for live web/UI work. |
| **Embedded Rust runtime** | The same crate that runs headless is loaded in-process via crate-type = ["cdylib", "staticlib", "rlib"]. The UI talks to it over a loopback WebSocket/HTTP gateway — no serialization-heavy IPC, no external daemon. |
| **Intent-routed coding modes** | Auto routes each turn to the best fit; or pin any of the 14 modes explicitly — **Agent** (default, full autonomy), **Vibe**, **Spec**, **Plan** (read-only plan authoring), **Ask** (read-only Q&A with citations), **TDD**, **Debug** (4-stage root-cause + in-app browser QA), **Architect**, **Pair**, **Context**, **MVAI**, **Harness**, **Curator** (research → DOCX / implementation blueprint), **Designer** (ten UI/design surfaces). Each mode rewires the system prompt, tool allowlist, approval policy, and auto-verify behavior. |
| **130+ tools** | File ops, PTY-mirrored shell, Git, ripgrep/content search, glob/multi-edit/patch-apply, **code intelligence** (tree-sitter outline + symbol graph for callers/implementors/uses + heuristic lexical code index), multi-engine web search/fetch, headless & embedded browser, SQLite + vector memory, todo/plan tracking, image generation, office docs (xlsx/pdf/docx), MCP tools, Skills, and Subagent delegation. |
| **Multi-provider** | OpenAI-compatible (incl. DeepSeek / Gemini-compatible endpoints), Anthropic, OpenRouter, GitHub Copilot, Claude Code, Ollama, Azure OpenAI, AWS Bedrock, Telnyx, and local CLI bridges. Provider keys and per-session model routing are configured in-app. |
| **Persistent memory & checkpoints** | SQLite + Markdown backends with vector embeddings, per-session work-dir isolation, and rewind/restore checkpoints. Interrupted turns are repaired and resumed against the *most recent* task on "continue". |
| **Automations & extensibility** | Cron-driven automations, Hooks, user Rules, Skills, MCP servers, multi-channel adapters (Slack / Telegram / Discord / Lark / … plus feature-gated Matrix), and first-party TypeScript & Python SDKs over the JSON-RPC 2.0 server. |
| **Performance-tuned** | Virtualized message list, rAF-coalesced streaming flushes, prioritized WebSocket heartbeats, content-aware context compaction, and spawn_blocking on IO hot paths. Hard constraints live in AGENTS.md. |
---
The crate is feature-gated so headless and desktop builds only compile what they need.
| Flag group | Examples | Notes |
|---|---|---|
| **default** | observability-prometheus, skill-creation, fs-watch, sandbox, lsp-push-diagnostics, tool-image, tool-utility-misc, tool-search-broad, tool-workspace-deep, tool-curator, lan-comms, office-docs | What a normal build ships. |
| **code-intel** | tree-sitter grammars (rust/js/ts/python/go/java/c/cpp/…) | Enables AST outline + symbol graph; falls back to heuristics when off. |
| **extras** | tool-cron, tool-sop, tool-team, tool-reports, tool-cloud-ops, computer-use, … | Optional tool families (also what desktop enables). |
| **channels** | channel-slack, channel-telegram, channel-matrix, channel-lark, … | Multi-channel adapters. |
libwebkit2gtk-4.1-dev libgtk-3-dev libayatana-appindicator3-dev librsvg2-dev patchelf.Pre-built installers are published on the Releases page. Every installer ships both the desktop app and the sen CLI so a single install gives you the GUI and a terminal-callable sen command.
| Platform | Asset | What lands on your system |
|---|---|---|
| **Windows x64** | SenWeaverCoding_<ver>_x64-setup.exe (NSIS) | Lets you pick install scope (per-user / per-machine) **and the install directory**. Adds the install dir to HKCU\Environment\Path so sen.exe is callable from any cmd / PowerShell / Windows Terminal session immediately after install. Removed cleanly on uninstall. |
| **Windows x64** | SenWeaverCoding_<ver>_x64_en-US.msi | For silent / group-policy deployment. |
| **macOS (universal)** | SenWeaverCoding_<ver>_universal.dmg | One drag-to-Applications package that runs natively on both Apple Silicon and Intel Macs (the bundled sen is a lipo-merged fat binary). CLI lives at SenWeaverCoding.app/Contents/Resources/sen; symlink it once with ln -sf "/Applications/SenWeaverCoding.app/Contents/Resources/sen" /usr/local/bin/sen to call it from any shell. |
| **Linux x64** | SenWeaverCoding_<ver>_amd64.deb | sudo dpkg -i on Debian/Ubuntu — installs the desktop launcher **and** /usr/bin/sen automatically. |
| **Linux x64** | SenWeaverCoding_<ver>_amd64.AppImage | Portable single-file desktop app. The CLI is **not** carried by the AppImage — install the .deb instead if you want a system-wide sen command. |
After installation launch SenWeaverCoding from the start menu / Applications / app launcher. On first run open Settings → Providers and add an API key for any provider you want to use. From a terminal you can verify the CLI with sen --help.
---
cd desktop
bun install
bun run tauri build # all bundlers configured in tauri.conf.json
The artifacts land in desktop/src-tauri/target/release/bundle/:
| Bundle | Path |
|---|---|
| Windows NSIS / MSI | bundle/nsis/*.exe, bundle/msi/*.msi |
| macOS dmg / .app | bundle/dmg/*.dmg, bundle/macos/*.app |
| Linux deb / AppImage | bundle/deb/*.deb, bundle/appimage/*.AppImage |
高质量的开源AI工作流项目,值得关注
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
✅ MIT 协议 — 最宽松的开源协议之一,可自由商用、修改、分发,仅需保留版权声明。
AI Skill Hub 点评:智能AI工作流 的核心功能完整,质量优秀。对于自动化工程师和运维人员来说,这是一个值得纳入个人工具库的选择。建议先在非生产环境试用,再逐步推广。
| 原始名称 | SenWeaverCoding |
| 原始描述 | 开源AI工作流:A Rust-first autonomous AI agent runtime and CLI code editor. Built on SenAgentO。⭐13 · Rust |
| Topics | aiclicoding-assistantsdesktop-apprust |
| GitHub | https://github.com/senweaver/SenWeaverCoding |
| License | MIT |
| 语言 | Rust |
收录时间:2026-06-25 · 更新时间:2026-06-26 · License:MIT · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型,复制安装指令后粘贴到对应客户端