spark-agent MCP Server
alexanderizh/spark-agent
[](https://agentrank-ai.com/tool/alexanderizh--spark-agent/?utm_source=badge&utm_medium=readme&utm_campaign=agentrank_badge) SparkWork is a local-first desktop Agent workbench for professional development and complex work. Orchestrate workflows, collaborate across Agent sessions, route tasks to the right models, and extend the workspace with custom tools and apps. Reviewable, steerable, reversible runs on macOS, Windows and Linux.
claude mcp add agentrank -- npx -y agentrank-mcp-server Overview
alexanderizh/spark-agent is a TypeScript tool. SparkWork is a local-first desktop Agent workbench for professional development and complex work. Orchestrate workflows, collaborate across Agent sessions, route tasks to the right models, and extend the workspace with custom tools and apps. Reviewable, steerable, reversible runs on macOS, Windows and Linux. Topics: ai-agent, claude, codex, desktop-app, electron, local-first, mcp, workflow-automation, ai-assistant, ai-workbench, claude-code, generative-ai, knowledge-base, llm, rag, spark-work, agent-workflow, developer-tools, multi-agent.
Ranked #5660 out of 24846 indexed tools.
Ecosystem
Score Breakdown
103 stars → growing community
Last commit 29d ago → recently updated
1/4 issues closed → many open issues
4 contributors → small team
No dependents → no downstream usage
Weights: Freshness 20% · Issue Health 20% · Dependents 22% · Stars 10% · Contributors 8% · How we score →
How to Improve
Matched Queries
From the README
<div align="center"> # SparkWork **本地优先的桌面 AI Agent 工作台:在一个应用里完成对话、开发、调研、文档、多媒体创作与自动化。** [官网](https://www.yiqibyte.com/) · [下载](#下载安装) · [能力图解](#能力图解) · [功能特性](#功能特性) · [快速开始](#快速开始) · [从源码构建](#从源码构建) · [工程手册](https://alexanderizh.github.io/spark-agent/) · [贡献指南](CONTRIBUTING.md) · [更新日志](CHANGELOG.md) </div> ## ✨ 项目简介 SparkWork(仓库名 Spark Agent)是一个运行在 macOS、Windows 和 Linux 上的桌面 AI Agent 工作台。 > 项目仍在快速迭代,部分功能和界面可能继续调整。 ## 📷 界面预览 ## 能力图解 下面八张图对应八条能力线的主链路;逐项细节见[官网文档中心](https://www.yiqibyte.com/docs)与[工程手册](https://alexanderizh.github.io/spark-agent/capabilities.html)。 <table> <tr> <td width="50%" valign="top"> **长期记忆:一轮对话的写入与读取闭环** 每轮结束异步抽取写入、每轮开始混合检索注入;写入过五道闸门,读取走 FTS5 与向量双路、RRF 融合后按 token 预算裁剪。 </td> <td width="50%" valign="top"> **知识库:从生成到读取的架构** 三条生成管道(对话沉淀 / 技能提议 / Repo Wiki)都先出草案、再人工确认,写入收敛到同一原语;读取零预注入,人机同源,先检索、再按需取正文。 </td> </tr> <tr> <td width="50%" valign="top"> **工作流:一张图的真实执行** 13 种节点类型经拓扑调度逐个真实派发:agent 节点是同一会话内的成员 turn,工具节点可配置成不经 LLM 的确定性直调;快照落库支持断点续跑。 </td>Read full README on GitHub →
Get the weekly AgentRank digest
Top movers, new tools, ecosystem insights — straight to your inbox.