多模型视觉理解 MCP 服务器,为不支持图片理解的 AI 编码模型提供视觉能力:分析截图、报错、UI 与文档,可接入多家主流视觉大模型。Multi-model vision MCP server that adds image understanding to AI coding models without native vision — analyze screenshots, errors, UI and documents via major vision LLM providers.
Capture a logged-in browser task once — replay it forever at zero LLM tokens. Local-first browser-automation MCP for Claude Code, Cursor & any MCP host; credentials never leave your machine.
Go-only, report-first code scanner exposed as an MCP server. Indexes a module, runs deterministic analyzers, and reports genuine bugs, dead code, stale docs & outdated APIs — never editing your code until you ask.
Geocode places, fetch global weather forecasts, ERA5 historical climate, marine conditions, air quality, and terrain elevation via MCP. STDIO or Streamable HTTP.
Persistent, verified memory for coding agents — so they stop re-explaining your codebase and never act on stale knowledge. Every memory is checked against your actual code; lives in your repo as plain files, shared via git. No account, no DB. Install: npx -y @kage-core/kage-graph-mcp install
AI writes. SPARDA proves. A deterministic, offline gate that catches when an AI edit removes a guard, exposes a route, or breaks an invariant — no API key, right in the agent edit loop.
Local-first MCP server to query and maintain a Markdown vault from Claude, Cursor, Gemini, Codex & other AI clients - no cloud, no context dump. Guided setup + standalone binaries.
Inspectable, source-traceable public-equity research for Codex, Claude Code, OpenCode, Grok Build and ChatGPT Work developer mode. 26 provisional method lenses, 34 MCP tools. Research software, not investment advice.
Decentralized networking protocol for AI agents — sovereign identity, peer-to-peer capability discovery, and direct encrypted connections. Works as an MCP server.