# pline.ai [Health: Active]

**Category:** 💻 Developer Tools  
**Repository:** https://github.com/grepsr/pline-api  
**GitHub Stars:** 1  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/pline-ai

## Description
Scrape, crawl, batch-scrape, map, and search the web through the hosted pline.ai API.

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

```json
"mcpServers": {
  "pline-ai": {
    "command": "npx",
    "args": ["-y","pline-api"],
    "env": {
      "PLINE_API_KEY": "",
      "PLINE_BASE_URL": ""
    }
  }
}
```

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

## Documentation

## What pline.ai MCP server does

The pline.ai MCP server connects MCP clients to the hosted pline.ai scraping API. It provides ten tools for web data collection and discovery, covering individual page scraping, batch scraping, site crawling, site mapping, and web search. The scraping backend is hosted by pline.ai and is not included in this repository.

The tools are suitable for agents that need to retrieve web content or begin a broader collection task without implementing direct API calls themselves. The repository also contains an agent skill with instructions and a Python client, but that skill is separate from the MCP server.

## How it works

For local use, an MCP client starts the `pline-api` npm package as a stdio subprocess. The server sends requests to the configured pline.ai API base URL and authenticates them with the supplied API key. The default base URL is `https://apix.pline.ai/v1`.

Remote agents can use the Streamable HTTP transport instead. The hosted connection uses Pline OAuth: users sign in, select a workspace API key, and connect through the `/mcp` URL. The repository includes separate documentation for hosting and OAuth configuration.

The included agent skill can use the connected MCP tools when they are available. If they are not available, it can run the bundled Python client instead. That skill requires Python 3.10 or newer, while the MCP package requires Node.js 18 or a supported alternative such as Docker.

## Setup and configuration

The quickest installation is:

```sh
npx -y pline-api@latest init
```

The initializer detects supported clients, requests an API key, and adds the MCP configuration. It can target specific clients, preview changes with `--dry-run`, or run non-interactively with `--yes`. Existing configuration is merged rather than replaced, and backups are created before changes are written.

For manual setup, configure the client to run `npx -y pline-api` and provide:

```json
{
  "PLINE_BASE_URL": "https://apix.pline.ai/v1",
  "PLINE_API_KEY": "your-own-api-key"
}
```

`PLINE_API_KEY` identifies the account or workspace used for API requests. `PLINE_BASE_URL` can point to another API base URL. The server does not load a local `.env` file automatically; expose these variables through the MCP client, shell, container, or hosting environment.

## Tools and capabilities

The pline.ai MCP server provides tools for:

- Scraping individual web pages
- Submitting batch scrape work
- Crawling websites
- Mapping website structure
- Searching the web
- Using the remaining tools exposed by the pline.ai API integration

The exact request parameters and response behavior are defined by the embedded API reference and the pline.ai service. An API key and network access to the configured endpoint are required for useful operation.

## Limitations and notes

This repository contains the integrations, not the scraping backend. Running `npx -y pline-api` directly starts a server that waits for an MCP client; it is not the interactive installer. Run `init` for automated client configuration.

The local MCP package requires Node.js 18 or newer unless using Docker, prebuilt binaries, or a source build. Client configuration must be restarted after installation or updates. Removing the integration requires deleting its client configuration entry; the initializer does not automatically remove every installation.

The separate agent skill has different requirements from the MCP server: it needs Git and Python 3.10 or newer. It is installed through `npx skills add grepsr/pline-api` and is not published as an npm skill package.

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