Klever MCPs

The Model Context Protocol (MCP) is an open standard that lets AI assistants (Claude Code, Claude Desktop, Cursor, VS Code, Windsurf, ChatGPT, and more) connect to external data and tools. Klever publishes a set of MCP servers so your AI assistant can write, deploy, query, and debug Klever-aware code without leaving your editor.

Available Klever MCPs

MCP	Purpose	Docs
Klever VM MCP	Smart-contract knowledge, scaffolding, and deployment helpers for the Klever VM (Rust / WASM).	Klever VM →
Klever Connect MCP	Knowledge / RAG MCP for the Klever Connect SDK — query and search Klever JavaScript/TypeScript SDK documentation, examples, and best practices.	Klever Connect →

Looking for a one-page setup overview? The ai.klever.org landing page has copy-paste install snippets for every supported client.

How to choose

• Building smart contracts on the Klever VM (Rust contracts, deploy, debug)? Use Klever VM MCP.
• Building dApps, wallet integrations, or backend services with @klever/connect-* JS/TS packages? Use Klever Connect MCP.
• Both can be installed side-by-side — the marketplace plugin flow keeps them isolated and you can enable either one per workspace.

Install paths (all MCPs)

Each MCP supports the same three install paths. See the per-MCP Setup page for verbatim snippets.

1. Klever Plugins Marketplace (recommended) — /plugin marketplace add klever-io/klever-marketplace, then /plugin install <slug>@klever-marketplace. Works in Claude Code and any client that speaks the marketplace plugin spec.
2. Manual install — run the published npm package directly via npx, or clone + build the source repo, and point your client's MCP config at it.
3. Public hosted endpoint — point your client at the hosted HTTPS MCP URL (e.g. https://mcp.klever.org/mcp for VM, https://mcp.klever.org/sdk-js for Connect).

Next steps

• Klever VM MCP →
• Klever Connect MCP →
• AI Development on Klever (overview)