UI Debugger MCP server

UI Debugger MCP

An MCP server that debugs UIs autonomously — so the AI that wrote your app can also test it, without a human clicking through every flow.

The problem

AI coding agents (Claude, etc.) are great at writing code. They're bad at knowing if the UI actually works. For backend code there are unit and integration tests. For UI, a human still has to open the app, log in, click around, and report what's broken. That human-in-the-loop is slow, boring, and the main bottleneck when an entire product is built by AI.

The idea

Eliminate the human from the UI-debug loop with an MCP server.

the browser or desktop, watches console + network, takes screenshots.

• A smart agent (Claude Code, Cursor, …) finishes a PR and wants to verify the UI.
• It hands a story to this server: "on web, log in and do X, Y, Z — tell me if it breaks."
• A small fast agent runs inside this server (via the Vercel AI SDK). It drives

• It reports structured findings back: pass/fail, what broke, evidence.
• The smart agent fixes the code and asks again. Loop until the UI works.

Unlike playwright-mcp — where the smart model issues every single click itself — here the smart model stays high-level and delegates the whole clicking loop to the small agent.

How it's different from playwright-mcp

Architecture — the three actors

Picture a boss, a fast blind driver, and a describer with eyes:

┌─────────────┐   MCP conversation    ┌──────────────────────────────────────┐
   │ smart agent │  start_debug ───────▶ │        UI Debugger MCP server         │
   │  (Claude)   │  send_message (live)  │                                       │
   │             │ ◀─────── get_findings │   ┌────────────┐     ┌────────────┐   │
   │ sets goals  │                       │   │  fast guy  │ look│ vision guy │   │
   │ fixes code  │                       │   │  (driver)  │────▶│  (eyes)    │   │
   │ loops       │                       │   │ deepseek   │◀────│  glm 5v    │   │
   └─────────────┘                       │   │ text·blind │ desc│ image      │   │
          ▲                              │   └─────┬──────┘     └────────────┘   │
          │ "works + looks nice"         │     observe / act (SQL-like)          │
          │ findings + screenshots       │         │ shared adapter contract     │
          └──────────────────────────────│─────────┼─────────────────────────────│
                                          └─────────┼─────────────────────────────┘
                                                    ▼
                              ┌──────────────┬──────────────┬─────────

the code**, loops. Stays high-level — never clicks.

• smart agent — the boss (Claude/caller). Sends a goal, reads findings, **fixes

click loop on structure (DOM / a11y tree / view hierarchy). Default: deepseek.

• fast guy — the driver. Fast, cheap, text-only and blind. Runs the

"does this look right? is the button centred?" and gets a description back. Default: glm. Spent only when visual judgment is needed.

• vision guy — the eyes. Multimodal. The driver calls look to ask

One goal: the UI works and looks nice. Full design in docs/idea/.

Shortened. The full README is on GitHub.