Skip to main content
MCP server CX Codex

Ollav — Claude Code MCP server: what it does, how to install

Ollav deploys HTML prototypes to live links and feeds element-level team feedback directly back into your coding agent.

MIT updated 2026-10-29

Install into Claude Code · Platform CLI: remote server All methods and platforms ↓
claude mcp add --transport http ollav https://app.ollav.app/mcp
Open on GitHub

What Ollav does

This tool turns static prototypes into interactive design reviews. When your agent finishes a layout, it runs share_html to upload the markup to a live URL. Teammates can then click directly on elements and leave inline notes.

Your agent inspects incoming feedback threads using get_review, applies the requested edits and calls publish_new_version. The public address stays identical. Once the layout matches expectations, the agent closes finished items with resolve_comment.

Hosting is handled entirely on the web at https://app.ollav.app/mcp. Without an active ol_test_... or ol_live_... token stored in OLLAV_API_KEY, all remote endpoint calls fail instantly.

What it helps with

  • share HTML mockups with teammates through temporary links
  • retrieve review threads linked to explicit DOM nodes
  • push layout revisions without changing the preview URL
  • mark addressed review items as resolved from the terminal

Example prompts for the agent

  • Upload preview.html to Ollav and print the public link
  • Check open feedback on the current prototype using get_review
  • Publish the modified HTML as a new version and resolve finished threads

How to install Ollav

Platform CLI: remote server

claude mcp add --transport http ollav https://app.ollav.app/mcp

Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more

Questions about Ollav

How to install Ollav in Claude Code?

Run claude mcp add --transport http ollav https://app.ollav.app/mcp in your command line.

What does Ollav do?

It lets coding assistants host HTML files online, inspect feedback pinned to specific elements, and push revised versions to the same link.

Do I need to run a local daemon?

No local binary is needed. Your assistant connects via HTTP directly to https://app.ollav.app/mcp using your account secret key.

Similar MCP servers

All MCP servers for Frontend and UI →

Project sponsors

Companies that keep this analytics open