# codify-mcp [Health: Active]

**Category:** 📂 Browser Automation  
**Repository:** https://github.com/cybairfly/codify-mcp  
**GitHub Stars:** 0  
**npm Downloads (last month):** 146  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/codify-mcp

## Description
Easily create reusable automation tools from casual browser actions using Apify Agent

## Tools
Capabilities this server exposes over MCP:

- **script** — Playwright automation code. Receives `inputs` object with user-provided parameters and `page` object for browser automation.
- **actorId** — Apify actor to execute. Defaults to `cyberfly/apify-agent`.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "codify-mcp": {
    "command": "npx",
    "args": ["-y","codify-mcp"],
    "env": {
      "APIFY_TOKEN": ""
    }
  }
}
```

**Requires environment variables:** `APIFY_TOKEN` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What codify-mcp MCP server does

The codify-mcp MCP server turns user-defined browser workflows into MCP tools that an AI assistant can call. Each tool has a name, description, input schema, and implementation. The implementation runs JavaScript Playwright code through an Apify actor, allowing the workflow to navigate pages, inspect content, and return a value or object.

Tool definitions are supplied as JSON strings on the command line. You can start one tool for development or pass several definitions to expose a larger collection. The project is intended for reusable browser tasks such as extracting information from a page, provided the required automation logic is written in the tool definition.

## How it works

When the process starts, it reads the JSON arguments, registers valid definitions with the MCP server, and makes them available to the connected assistant. During execution, the script receives an `inputs` object containing the caller's parameters and a `page` object for Playwright browser automation.

The default actor is `cyberfly/apify-agent`. A definition can set another `actorId` when a different Apify actor should execute the script. The script must be valid JavaScript and should await browser operations before returning its result.

The codify-mcp MCP server is stateless: each invocation is independent. Tool definitions remain in the command arguments rather than separate configuration files. Once installed, the bundled dependencies can run locally without downloading additional dependencies.

## Setup and configuration

Run the package directly with `npx codify-mcp` or install it globally with npm. Append each tool definition as a JSON argument. Multiple arguments create multiple exposed tools.

Apify authentication is required for tool execution. The server first checks `APIFY_TOKEN`, then checks `~/.apify/auth.json`, which can be created with `apify login`. If neither source provides a token, execution fails. The token can also be placed in an MCP client's environment configuration.

Claude Desktop configuration uses an MCP server entry with `command: "npx"`, the package name in `args`, and optionally `APIFY_TOKEN` in `env`. The README also names Cursor and VS Code as connection targets.

## Tools and capabilities

- Define MCP tools with names, descriptions, and JSON input schemas.
- Run Playwright scripts with access to `inputs` and `page`.
- Select an Apify actor with `actorId`, or use `cyberfly/apify-agent` by default.
- Pass one or many tool definitions through command-line arguments.
- Support schema fields such as required values, defaults, integer types, and enumerations.
- Return script-generated values or objects to the calling assistant.

## Limitations and notes

The codify-mcp MCP server does not generate the automation script itself. The supplied JavaScript must be valid, use the provided objects correctly, and return a result. Missing or malformed JSON arguments produce tool-registration errors, while missing Apify authentication prevents execution.

For collections of 50 or more tools, the README recommends splitting them across multiple MCP server processes. Authentication sessions through the project's separate Login component are described as a future item, so this listing does not establish that feature as available.

_Full upstream README: https://allmcps.com/mcp/codify-mcp/readme_