harvis-dev-mcp
Remote MCP server for harvis.dev — deploy static sites from AI agents and get a live URL in seconds. Endpoint: https://harvis.dev/api/mcp
Documentation
harvis.dev MCP Server
Remote MCP server for harvis.dev — static site hosting for AI agents. Hand it a folder of files from any MCP client and get a live public URL back in seconds. No account, no API key.
https://harvis.dev/api/mcp- Transport: Streamable HTTP (stateless)
- Auth: none — deploys are anonymous; each deploy returns a private claim link so a human can take ownership of the site later
- Registry: `dev.harvis/harvis` in the official MCP Registry · Glama
> This repository documents the hosted server. There is nothing to install or run from here — the server lives at the URL above and this repo carries its manifest, client configuration, and tool reference.
Quickstart
Claude Code
claude mcp add --transport http harvis https://harvis.dev/api/mcpClaude Desktop, Cursor, VS Code, and other MCP clients
Add a remote (Streamable HTTP) server with URL `https://harvis.dev/api/mcp`. Typical JSON config:
{
"mcpServers": {
"harvis": {
"type": "http",
"url": "https://harvis.dev/api/mcp"
}
}
}For clients that only speak stdio, bridge with `mcp-remote`:
{
"mcpServers": {
"harvis": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://harvis.dev/api/mcp"]
}
}
}Tools
`deploy_site`
Publish a static website and get a live public URL back instantly.
| Input | Type | Description |
|---|---|---|
| `files` | array (required) | Every file of the site: `{ path, content, encoding? }`. Paths are relative to the site root (include an `index.html`); `content` is UTF-8 text, or base64 with `encoding: "base64"` for binary files. Up to 50 MB total. |
| `name` | string | Optional human-friendly site name. |
| `subdomain` + `deployToken` | string | Pass both, from an earlier deploy, to update that site in place instead of creating a new one. |
Returns the live site `url` (`https://{slug}.harvis.dev`), a private `claimUrl` (open it and sign in — free — to attach the site to an account for later edits; it cannot be recovered if lost), the `subdomain`, and a private `deployToken` for future in-place updates.
Try it from the terminal
curl -X POST https://harvis.dev/api/mcp \
-H 'Content-Type: application/json' \
-H 'Accept: application/json, text/event-stream' \
-d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-03-26","capabilities":{},"clientInfo":{"name":"curl","version":"1.0"}}}'Or run the official inspector:
npx @modelcontextprotocol/inspector --cli https://harvis.dev/api/mcp --transport http --method tools/listOther ways to deploy to harvis.dev
The MCP server is one of several deploy paths — there is also a CLI (`npx harvis`), browser drag-and-drop, an HTTP API, and a dashboard. See harvis.dev for all of them.
Support
- Website: https://harvis.dev
- Issues with the MCP server: open an issue here
- Email: dev@harvis.dev
Frequently asked questions
What is harvis-dev-mcp?
harvis-dev-mcp is Remote MCP server for harvis.dev — deploy static sites from AI agents and get a live URL in seconds. Endpoint: https://harvis.dev/api/mcp
How do I install harvis-dev-mcp?
Open the GitHub repository and follow its README. Most MCP servers are added to your client's MCP config, then called by your agent.
Is harvis-dev-mcp open source?
Yes — it is hosted on GitHub at https://github.com/harvis-io/harvis-dev-mcp.
Related MCP tools
Model Context Protocol Servers
Pre-indexed code knowledge graph, auto syncs on code changes, for Claude Code, Codex, Gemini, Cursor, OpenCode, AntiGravity, Kiro, CoPilot, and Hermes Agent — fewer tokens, fewer tool calls, 100% local
an open source, extensible AI agent that goes beyond code suggestions - install, execute, edit, and test with any LLM
The easy-to-use open source Business Intelligence and Embedded Analytics tool that lets everyone work with data :bar_chart:
The Open-Source Multimodal AI Agent Stack: Connecting Cutting-Edge AI Models and Agent Infra
Cognee is the open-source AI memory platform for agents. Give your AI agents persistent long-term memory across sessions with a self-hosted knowledge graph engine.
Run your own MCP server? See who uses it and what to fix.
Measure it with TrackMCP