webmachinelearning/webmcp: 🤖 WebMCP
🤖 WebMCP. Contribute to webmachinelearning/webmcp development by creating an account on GitHub.
Welcome! Type "help" for available commands.
$
Developer Tool
Summary
Convex 1.19.5 introduces a beta MCP server that lets AI agents inspect deployments, run functions, and read or write data in Convex projects via the Model Context Protocol. The article walks through setup in Cursor, supported tools, and example use cases such as agents writing safe one-off JavaScript queries.
Highlights
- Beta MCP server included in Convex 1.19.5 for AI agent integration
- Supports introspecting deployments, tables, functions, and reading/writing data
- Agents can safely write sandboxed one-off queries that cannot mutate the database
- Setup instructions provided for Cursor via mcp.json and the npx convex@latest mcp start command
- Currently tested on Cursor with macOS Sequoia; other agents like Windsurf and Claude Desktop expected to work
auto-generated
Context
Audience
Backend developers using Convex who want AI agents to interact with their deployments
DomainBackend-as-a-Service and AI Agent Tooling
FormatTechnical blog post
AccessOpen beta via npm/Convex CLI
Topics
Related
Cursor IDEConvexModel Context ProtocolClaude DesktopWindsurf
