Skip to content
MCP server

Tavero

Finds UI drift in React apps and hands your coding agent each fix

Runs locallyStatus and history
Install
Pricing
Open source
Platforms
CLI, macOS, Windows
Runs
Locally
Version
0.7.0
Listing updated
2 days ago

About Tavero

Tavero is an open-source MCP server for developer tools. It runs locally on your machine and is started by your AI client.

Tavero works from a running dev server rather than from source alone. It opens the app in Chromium at mobile, tablet and desktop widths (plus one dark pass if a dark theme exists), maps every visible element back to the React component and source line that rendered it using React's dev-mode data, and adds nothing to your project. It writes only to a .tavero folder.

Four rules compare what renders against your design tokens and against itself: buttons, cards (marked preliminary), colors and scales. Each finding becomes a task with the file and line, the current value and the token or component expected, so an agent can edit, rescan a single route and confirm the finding is resolved.

The MCP server runs on stdio with nine tools, and the README prints ready-made config for Claude Code, Cursor, Codex, Gemini, VS Code, Windsurf and Antigravity. A local panel lets a person mark findings as fix, intended or canonical, and the agent sees those decisions on its next call. It needs React 19 in dev mode on Vite or Next; React 18 is not supported yet. Everything runs locally with no account and no telemetry.

Install

Run in a terminal

claude mcp add tavero -- npx -y tavero

Add to claude_desktop_config.json

{
  "mcpServers": {
    "tavero": {
      "command": "npx",
      "args": [
        "-y",
        "tavero"
      ]
    }
  }
}
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "tavero": {
      "command": "npx",
      "args": [
        "-y",
        "tavero"
      ]
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "tavero": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "tavero"
      ]
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "tavero": {
      "command": "npx",
      "args": [
        "-y",
        "tavero"
      ]
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "tavero": {
      "command": "npx",
      "args": [
        "-y",
        "tavero"
      ]
    }
  }
}

Tools this server provides (9)

Taken from the project documentation; the running server may differ.

tavero_context
Design-system brief to call before writing UI.
tavero_findings
Lists problems, worst first.
tavero_finding
Shows one problem with evidence, expected value and the exact rescan call.
tavero_scan
Scans the app, or only touched routes and rules after a fix.
tavero_diff
Shows resolved, new and unchanged findings since the previous scan.
tavero_components
Lists the component inventory.
tavero_component
Shows one component.
tavero_show
Shows crop paths for findings.
tavero_tokens
Lists design tokens.

What it does

Runtime UI scan
Scans the dev server at three widths and maps elements to components and source lines.
Four drift rules
Checks buttons, cards, colors and scales against your tokens.
Fix tasks for agents
Each finding gives file:line, current value and the expected token.
Rescan and diff
Rescans touched routes only and reports resolved, new and unchanged findings.
Local decision panel
Mark findings as fix, intended or canonical so the agent sees your choices.
Context brief
Produces a roughly 1,500-token summary of the design system for agents.

Good for

  • Catch hardcoded colors that are not design tokens
  • Find near-duplicate button styles across pages
  • Let a coding agent fix spacing and radius inconsistencies with exact edits
  • Review UI consistency before shipping a React app

Pricing

Open source

Runs locally; no account required

Prices change. Check taveroui.com before you buy.

Tavero FAQ

Which frameworks does it support?

React 19 in dev mode on Vite or Next (App Router, Turbopack). React 18 is not supported yet.

Does it send data anywhere?

No. The README says it runs locally, sends nothing anywhere, and has no account or telemetry.

How do I add it to an agent?

Run npx tavero mcp --print-config with your client name, such as claude, cursor or codex, and paste the output into the client's config.

Where does it write files?

Only to a .tavero folder in the project, which includes decisions.json that you can commit.

Where do design tokens come from?

From CSS custom properties and Tailwind 3 or 4 themes.

Is Tavero free?

Yes. Tavero is open source, so you can use and self-host it for free.

How do I add Tavero to Claude Code?

Run this command in your terminal: claude mcp add tavero -- npx -y tavero

What are the best alternatives to Tavero?

The closest alternatives to Tavero are REA, HyperFrames by HeyGen, Codebase Memory, GitHub and Serena MCP: the IDE for your agent.

Alternatives to Tavero

  1. REA

    MCP server

    CLI and MCP server that lets coding agents reverse engineer apps and binaries

    Runs locally
    • Open source
    • 464K GitHub stars
    • Developer tools
  2. Hosted MCP connector for creating and rendering videos from HTML through AI chat

    • Remote
    • Freemium
    • 85.3K GitHub stars
    • Design and media
  3. Codebase Memory

    MCP server

    Local MCP server that builds a code knowledge graph for AI coding agents

    Runs locally
    • Open source
    • 48.7K GitHub stars
    • Knowledge and memory
  4. GitHub

    MCP server

    Official MCP server connecting AI assistants to GitHub repos, issues and workflows

    Not checked yet
    • Remote
    • Open source
    • 34.5K GitHub stars
    • Developer tools
  5. MCP toolkit giving coding agents symbol-level code retrieval and editing

    Runs locally
    • Open source
    • 31.5K GitHub stars
    • Developer tools
  6. XERJ

    MCP server

    Local search engine for AI agents with cited answers over your code and docs

    Runs locally
    • Open source
    • 19.2K GitHub stars
    • Knowledge and memory