Webvizio MCP Server
Webvizio/mcp
Converte feedback de usuários em tasks prontas para agentes IA resolver
Ver no GitHubMCP server que converte feedback e relatórios de bugs de websites em tarefas estruturadas para agentes IA. Integra-se com ferramentas de codificação (Cursor, Windsurf) fornecendo screenshots, logs e contexto completo para resolução automática de problemas.
Instalação
npx convez-ai Webvizio/mcp1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Automatizar correção de bugs relatados por usuários finais
- Integrar feedback visual com fluxo de desenvolvimento IA
- Centralizar tarefas de desenvolvimento com contexto de reprodução (screenshots, logs)
Veredito

Ferramenta bem pensada que fecha a lacuna entre feedback não-técnico e execução de código. Requer infraestrutura (extensão browser + API key) mas entrega valor tangível em automação de bug-fixing com agentes IA.
README
A tradução para o português deste README ainda está na fila.
About Webvizio
Webvizio streamlines communication between end-users, non-technical teams, and developers.
Our easy-to-use browser extension lets you capture visual feedback and bug reports directly on a web page. It then automatically gathers all relevant data and converts it into actionable, context-enriched development tasks for your AI coding agents.
From non-technical product feedback to bug-free code in seconds, not days. Learn more about Webvizio for developers here.
Webvizio MCP Server
A TypeScript-based Model Context Protocol (MCP) server that provides secure and structured access to the Webvizio API. This server enables MCP clients to seamlessly interact with Webvizio projects and development tasks through a standardized interface.
Webvizio MCP Server - Automatically converts feedback and bug reports from websites and web apps into actionable, context-enriched developer tasks. Delivered straight to your AI coding tools, the Webvizio MCP Server ensures your AI agent has all the data it needs to solve tasks with speed and accuracy.
📋 Available Tools
| Tool | Method | Description |
|---|---|---|
get_projects | GET | Fetch all available Webvizio projects |
get_current_project | GET | Get details of the currently selected project |
set_project | POST | Set the current Webvizio project to work with the project tasks |
get_tasks | GET | Get the list of open tasks assigned to a user |
get_task_description | GET | Get the task description |
get_task_prompt | GET | Get an AI prompt to execute the task |
get_task_error_logs | GET | Get the task error logs |
get_task_console_logs | GET | Get the task console logs |
get_task_network_logs | GET | Get the task network logs |
get_task_action_logs | GET | Get the task action logs (Repro steps) |
get_task_screenshot | GET | Get the task screenshot |
close_task | POST | Set task in Webvizio to Done status |
🛠 Installation
- Make sure you have Node.js 18 (or a higher version) installed on your device.
- Copy the MCP server configuration code to the
mcpServerssection of your AI Client configuration file."webvizio": { "command": "npx", "args": [ "-y", "@webvizio/mcp-server" ], "env": { "WEBVIZIO_API_KEY": "<your-api-key>" } } - Generate and copy the API key for the MCP server on the AI Settings page https://app.webvizio.com/profile/ai
- Insert the API key into the
WEBVIZIO_API_KEYenvironment variable. - Save the changes.
- Done. Your MCP server is configured and ready to work.
🚀 Usage
- Open your IDE's chat (e.g., agent mode in Cursor or Cascade in Windsurf).
- Paste a link to a Webvizio task. Alternatively, you can ask the AI agent to display all your tasks.
- Instruct the AI agent to execute the selected task.
- The AI agent can request the use of tools to gather more information on the task (e.g., screenshots, console logs, etc.). Approve these requests to allow the agent to proceed.
- Check the agent's work to ensure the task has been completed correctly.
- When the task is complete, ask the AI Agent to close the task.