model-context

Talk with your notes in Claude. RAG over your Apple Notes using Model Context Protocol.

Coral is an open-source, local-first session bridge that lets AI agents borrow a user's already-authenticated browser sessions on a per-site, per-action, fully audited basis.

Make your website agent-ready: a coding-agent plugin for Claude Code and Codex that adds WebMCP tools to your web app

🚀 AI-powered freelance gig aggregator MCP server for Claude Desktop. Search gigs, generate proposals, negotiate rates & optimize profiles with ChatGroq AI.

Generate safe, typed, human-reviewed WebMCP tools from the API contracts you already have, instead of hand-writing registerTool() calls for every action.

Make your website agent-ready with WebMCP. Coding-agent plugin for Claude Code, Codex, and Cursor: audit, implement, and verify document.modelContext tools.

DERELICT — an asymmetric human+AI co-op escape room. You play the ship; your AI agent runs its systems through 31 live WebMCP tools. Three chapters, four joint-ritual endings, seeded ships where every run is unique, EN/pt-BR.

A pixel-art canvas a person and an AI agent draw on together. Six WebMCP site tools. Built for OpenAI's WebMCP Challenge.

CLI based MCP client.

Expose a page's WebMCP tools as a real MCP server: an in-page JSON-RPC handler (initialize/tools/list/tools/call) over a pluggable transport.

Client-side human-in-the-loop for WebMCP: gate agent tool calls behind an approval policy (auto/confirm/deny), with a built-in modal or your own approver.

Zero-config highlighting of the DOM changes an AI agent's WebMCP tool calls make — overlay rings, change timeline, off-screen reveal. Framework-agnostic core + React binding.

Measure WebMCP providers across models and runtimes. Attribute every failure to the layer that caused it. Keep the provider blameless when it's not at fault.

Define, validate, and register structured browser tools for the experimental WebMCP API

Quick-reference guide to understanding common tech protocols — what they do and how to choose the right ones for tasks like secure logins, file transfer, video output, and building AI workflows. Includes comparisons of web, device, and model interaction protocols.

Model-agnostic runtime for executing stateful, multi-step trajectories across independent WebMCP providers.

A neutral execution-control seam for agent tool calls: allow / block / indeterminate, only allow proceeds.

Zero-config dev panel for WebMCP: lists a page's tools, auto-builds a form from each inputSchema, invokes them, shows results. Pairs with webmcp-highlight.

Test WebMCP tools without a flagged Chrome or a live model: a mock navigator.modelContext plus a driver to list, validate-args, invoke, and snapshot the toolset.

Qualify WebMCP interoperability against frozen profiles: qualified / not_qualified / insufficient_evidence - no execution.