AIHackr

Read public webpages, extract structured data, discover links, and run bounded crawls.

GitHub

Hosted MCP Server

npx add-mcp 'https://aihackr.com/mcp'

Installs into Claude Code, Codex, Cursor and more

Documentation

AIHackr MCP

Read public webpages as Markdown or HTML, discover links, extract structured JSON, and run bounded crawls from your AI client.

Connect with OAuth

  1. Add https://aihackr.com/mcp as a remote MCP server using Streamable HTTP.
  2. Choose OAuth if prompted. Sign in to AIHackr with your email and review the permissions before authorizing the connection.
  3. Once connected, try the example below. You can revoke the connection from Connected apps.

If your client uses bearer authentication, create an API key in the AIHackr dashboard and set Authorization: Bearer YOUR_AIHACKR_API_KEY using the client's secure credential settings.

Client setup

  • ChatGPT · Add AIHackr as a custom MCP server with OAuth if your workspace supports custom MCP connections. MCP setup guide
  • Claude · In Customize → Connectors, add a custom connector named AIHackr and paste the endpoint.
  • Cursor · Add the configuration below to your MCP settings.
  • Gemini CLI · Run gemini extensions install https://github.com/biserd/aihackr-mcp, then use /mcp auth aihackr in Gemini CLI to connect your account.
  • VS Code · Install AIHackr in VS Code, or run MCP: Add Server, choose HTTP, and enter the endpoint. Then start the server from MCP: List Servers and follow the sign-in prompt.
{
  "mcpServers": {
    "aihackr": {
      "url": "https://aihackr.com/mcp"
    }
  }
}

Try one page

Use AIHackr to read https://fixtures.aihackr.com/ as Markdown. Return the page heading and source URL. Use one page.

The expected heading is “Native engine fixture.” A successful Markdown scrape uses one credit.

Plans and credits

Verified accounts receive 1,200 free credits per UTC calendar month. The playground, REST API, and MCP use the same workspace balance.

PlanUSD per monthCredits per month
Starter$73,000
Builder$157,500
Scale$2920,000

A successful scrape or link map uses one credit. Structured extraction uses five credits. A crawl reserves its requested page limit, charges one credit for each successful page with retained content, and releases unused credits when the job ends. Current plans · Usage and limits

MCP documentation · Privacy · Support

This repository contains the install configuration and listing manifest. The server implementation is in the AIHackr repository.