2Captcha MCP

Name: 2Captcha MCP
Author: 2Captcha

Web search, scraping, marketplace parsing and captcha solving through the 2Captcha service

Official · 2Captcha JavaScript Remote (hosted) · Streamable HTTP Local · stdio Browser automation Web search

GitHub npm Website

45 popularity score ★ 39 ↓ 146/wk

Connect your client

Remote (hosted)

Add in one click

Add to Cursor Add to VS Code Add to LM Studio Add to Goose

Opens the app, which asks before it adds the server.

Or add it by command or config

Claude Codeterminal

claude mcp add --transport http 2captcha https://mcp.2captcha.com/mcp \
  --header "Authorization: Bearer <API_TOKEN>"

qwen codeterminal

qwen mcp add --transport http 2captcha https://mcp.2captcha.com/mcp \
  --header "Authorization: Bearer <API_TOKEN>"

kimi code cli~/.kimi-code/mcp.json

{
  "mcpServers": {
    "2captcha": {
      "url": "https://mcp.2captcha.com/mcp",
      "headers": { "Authorization": "Bearer <API_TOKEN>" }
    }
  }
}

n8nMCP Client Tool node

Endpoint:       https://mcp.2captcha.com/mcp
Transport:      HTTP Streamable
Authentication: Header Auth
Header name:    Authorization
Header value:   Bearer <API_TOKEN>

Local (npm)

Add in one click

Add to Cursor Add to VS Code Add to LM Studio Add to Goose

Opens the app, which asks before it adds the server. Then replace each <NAME> in its settings with your value.

Or add it by command or config

It needs API_TOKEN. Replace each <NAME> with your value.

Claude CodeTerminal

claude mcp add 2captcha -e API_TOKEN=<API_TOKEN> -- npx @2captcha/mcp

Claude Desktopclaude_desktop_config.json

{
  "mcpServers": {
    "2captcha": {
      "command": "npx",
      "args": [
        "@2captcha/mcp"
      ],
      "env": {
        "API_TOKEN": "<API_TOKEN>"
      }
    }
  }
}

Cursor~/.cursor/mcp.json

{
  "mcpServers": {
    "2captcha": {
      "command": "npx",
      "args": [
        "@2captcha/mcp"
      ],
      "env": {
        "API_TOKEN": "<API_TOKEN>"
      }
    }
  }
}

VS Code.vscode/mcp.json

{
  "servers": {
    "2captcha": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "@2captcha/mcp"
      ],
      "env": {
        "API_TOKEN": "<API_TOKEN>"
      }
    }
  }
}

Codex~/.codex/config.toml

[mcp_servers.2captcha]
command = "npx"
args = ["@2captcha/mcp"]

[mcp_servers.2captcha.env]
API_TOKEN = "<API_TOKEN>"

Gemini CLITerminal

gemini mcp add --scope user -e API_TOKEN=<API_TOKEN> 2captcha npx -- @2captcha/mcp

Local (container)

Add in one click

Add to Cursor Add to VS Code Add to LM Studio Add to Goose

Opens the app, which asks before it adds the server. Then replace each <NAME> in its settings with your value.

Or add it by command or config

It needs API_TOKEN. Replace each <NAME> with your value.

Claude CodeTerminal

claude mcp add 2captcha -e API_TOKEN=<API_TOKEN> -- docker run --rm -i -e API_TOKEN ghcr.io/2captcha/2captcha-mcp:latest

Claude Desktopclaude_desktop_config.json

{
  "mcpServers": {
    "2captcha": {
      "command": "docker",
      "args": [
        "run",
        "--rm",
        "-i",
        "-e",
        "API_TOKEN",
        "ghcr.io/2captcha/2captcha-mcp:latest"
      ],
      "env": {
        "API_TOKEN": "<API_TOKEN>"
      }
    }
  }
}

Cursor~/.cursor/mcp.json

{
  "mcpServers": {
    "2captcha": {
      "command": "docker",
      "args": [
        "run",
        "--rm",
        "-i",
        "-e",
        "API_TOKEN",
        "ghcr.io/2captcha/2captcha-mcp:latest"
      ],
      "env": {
        "API_TOKEN": "<API_TOKEN>"
      }
    }
  }
}

VS Code.vscode/mcp.json

{
  "servers": {
    "2captcha": {
      "type": "stdio",
      "command": "docker",
      "args": [
        "run",
        "--rm",
        "-i",
        "-e",
        "API_TOKEN",
        "ghcr.io/2captcha/2captcha-mcp:latest"
      ],
      "env": {
        "API_TOKEN": "<API_TOKEN>"
      }
    }
  }
}

Codex~/.codex/config.toml

[mcp_servers.2captcha]
command = "docker"
args = ["run", "--rm", "-i", "-e", "API_TOKEN", "ghcr.io/2captcha/2captcha-mcp:latest"]

[mcp_servers.2captcha.env]
API_TOKEN = "<API_TOKEN>"

Gemini CLITerminal

gemini mcp add --scope user -e API_TOKEN=<API_TOKEN> 2captcha docker -- run --rm -i -e API_TOKEN ghcr.io/2captcha/2captcha-mcp:latest

Settings 9

`API_TOKEN`	Required: 2Captcha API key or an operator-issued bearer token; the server exits without it when MCP_URL is the default.
`GROUPS`	Optional: Comma-separated tool groups; default parsing,batch,captcha; browser, browser_full or all add more.
`TOOLS`	Optional: Comma-separated individual tool names to enable, combined with GROUPS.
`MCP_URL`	Optional: Remote endpoint to relay to; default https://mcp.2captcha.com/mcp.
`MAX_SPEND_USD`	Optional: Spend cap for the session, or per window such as 5/1h.
`RATE_LIMIT`	Optional: Client-side call limit such as 100/1h.
`MAX_CONCURRENCY`	Optional: Maximum tool calls in flight at once; unlimited by default.
`POLLING_TIMEOUT`	Optional: Per-call timeout in seconds; default 600.
`CACHE_DIR`	Optional: Where the offline tool-list cache is kept; default the OS cache directory.

In a container, a setting goes in with -e NAME=value after docker run.

What it does

2Captcha MCP is the captcha-solving company's own server for AI clients. It covers four jobs: searching the web, fetching pages as Markdown or HTML, turning shop product pages and listings into structured JSON, and solving captchas of the types 2Captcha supports. Batch tools run scraping or parsing over many URLs as background jobs, and an opt-in group drives a browser hosted by 2Captcha so an assistant can sign in past a captcha and keep the cookies.

The npm package does no scraping itself. Its server.js (v0.1.3) is a small stdio bridge: it fetches the tool list from the hosted service at https://mcp.2captcha.com/mcp, keeps it for 60 seconds and caches it on disk so the server still starts offline, and forwards every call with your token. The tools therefore change when the service changes, and the hosted endpoint offers the same set directly.

Only the parsing, batch and captcha groups (17 tools) are shown by default; GROUPS=browser adds 11 login tools and GROUPS=browser_full or all adds all 23 browser tools. The README counts 40 tools, but the bridge always adds a 41st, session_stats, that it answers locally (server.js). Two captcha tools, detect_captcha and solve_captcha_on_page, take HTML from a browser you already drive and return the JavaScript or cookie to apply there.

What it needs

To run it yourself	Node.js 18+orDocker

A 2Captcha account and its API key, set as API_TOKEN (sign-up is free and includes a monthly allowance; captcha solves are charged to the account balance). Node.js 18 or newer for npx, or Docker for the image. Network access to mcp.2captcha.com. The hosted endpoint needs an Authorization: Bearer header with the same key.

Limits

All work runs on 2Captcha's servers; the local package only relays calls, so nothing works offline beyond listing cached tools. The free allowance is 200 calls or $0.50 of measured spend per rolling 30 days per account; after that calls are refused until it refills. Captcha solves cost money per call, and extract always makes one model call. The hosted browser keeps one live page per account, so it is not suited to parallel work, and browser tools are off unless GROUPS includes them. MAX_SPEND_USD is checked against spend already reported, so it can overshoot by one call. DataDome and CaptchaFox answers need your own proxy and user_agent to be accepted in your session.

Who it suits

Suits developers who already hold, or are willing to open, a 2Captcha account and want scraping, search and captcha solving behind one key without running a browser themselves. A good fit for an assistant that drives its own browser and only needs help at a captcha, since the page-side captcha tools keep the session local. Pick a local browser server instead when pages must not be sent to a third party, when you need many browser pages at once, or when you want no per-call costs.

Tools 41

scrape_page: Fetches one URL and returns it as Markdown or raw HTML, rendering in a browser when needed
search_web: Runs a web search and returns ranked results with title, URL and snippet
discover_urls: Lists the URLs a site publishes in robots.txt and its XML sitemaps, with filters; does not crawl
discover_search_params: Reads one page to find the query parameters and form fields a site's own search uses
parse_marketplace: Turns a product page or a search listing into structured JSON such as price, rating and seller
extract: Pulls JSON matching a schema you supply out of a URL or pasted text, using one model call
get_account: Shows the account identity, enabled capabilities, spend counters and remaining free allowance
scrape_pages: Starts a background job that scrapes many URLs and returns a job id
parse_pages: Starts a background job that parses many marketplace URLs and returns a job id
get_job: Checks a background job's status and collects its per-URL results
cancel_job: Stops a running background job
list_captcha_types: Lists the captcha types that can be solved and the parameters each needs
solve_captcha: Solves one captcha of a supported type and returns the token or answer and its cost
detect_captcha: Identifies the captcha on a page from HTML you pass in, without solving it
solve_captcha_on_page: Detects and solves a page's captcha and returns JavaScript or a cookie to apply in your own browser
captcha_report: Marks a solution as good or bad; bad reports are refunded
captcha_balance: Returns the 2Captcha account balance in USD
browser_navigate: Opens a URL in the vendor-hosted browser
browser_click: Clicks an element in the hosted browser
browser_fill: Fills a form field in the hosted browser
browser_type: Types text into the focused element of the hosted browser
browser_press_key: Sends a key press to the hosted browser
browser_snapshot: Returns the page's accessibility tree with element references
browser_get_text: Returns the visible text of the current page
browser_get_html: Returns the HTML of the current page
browser_save_session: Stores the hosted browser's cookies on the vendor's server for later runs
browser_load_session: Restores a saved browser session
browser_list_sessions: Lists the saved browser sessions
browser_go_back: Goes back one page in history (browser_full group)
browser_go_forward: Goes forward one page in history (browser_full group)
browser_reload: Reloads the current page (browser_full group)
browser_scroll: Scrolls the page (browser_full group)
browser_select_option: Picks an option in a select element (browser_full group)
browser_hover: Moves the pointer over an element (browser_full group)
browser_drag: Drags one element onto another (browser_full group)
browser_console_messages: Returns the page's console output (browser_full group)
browser_evaluate: Runs JavaScript in the page (browser_full group)
browser_snapshot_items: Returns repeated items found on the page (browser_full group)
browser_screenshot: Takes a screenshot of the page (browser_full group)
browser_save_as_pdf: Saves the page as a PDF (browser_full group)
session_stats: Reports calls per tool and the rate-limit window for this session; answered locally

Popularity

45 popularity score, on our list (the bar is 36)

★ 39 ↓ 146/wk

GitHub stars	39	19.2 points
npm downloads, last week	146	17.3 points
Official server	Yes	8 points

Measured 2026-11-07. How the score works.

Where it is published

Code	2captcha/2captcha-mcp ★ 39
npm	@2captcha/mcp 0.1.3, 2026-10-12, the one the command runs
Container	ghcr.io/2captcha/2captcha-mcp
Hosted	`https://mcp.2captcha.com/mcp` Streamable HTTP · OAuth sign-in

What it can reach

Runs	On your machine, or on the maker's servers
Reaches	The internet, Drives a browser, Your account at a service
Changes things	Yes: it can write, send, create or delete
Sign-in	OAuth sign-in for the hosted endpoint; an API key or token to run it yourself
Licence	MIT (npm)
Last code change	2026-10-12

Code runs on 2Captcha's servers; the local package is a relay that sends each tool's arguments and your API key as a bearer token to MCP_URL (https://mcp.2captcha.com/mcp by default) and stores only a tool-list cache on disk. detect_captcha and solve_captcha_on_page send whatever page HTML you pass, which from a signed-in page includes that page's content. The opt-in browser groups drive a browser on the vendor's server, can click and fill forms on any site, and browser_save_session keeps that site's cookies server-side, so a saved session is a live login held by 2Captcha. solve_captcha and solve_captcha_on_page spend account balance on each call. Request data is retained under 2Captcha's privacy policy.

Over time

GitHub stars 39 +10 (+34 %)

3929

13 Oct 20267 Nov 2026

npm downloads a week 146 −150 (−51 %)

1.5k26

7 Sept 20267 Nov 2026

Before the site measured it, read from npm's daily download counts.

Popularity score 45 +2 (+4 %)

5140

13 Oct 20267 Nov 2026

What is rising across the directory

Sources

Page checked 2026-10-15.

Badge for your README

Shows this server's rank in its category, and stars for a server that runs on your machine. It updates every hour.

MarkdownREADME.md

[![2Captcha MCP on EveryMCPServer](https://everymcpserver.com/badge/2captcha-mcp-2.svg)](https://everymcpserver.com/servers/2captcha-mcp-2)

Connect your client

Add in one click

Or add it by command or config

Add in one click

Or add it by command or config

Add in one click

Or add it by command or config

Settings 9

What it does

What it needs

Limits

Who it suits

Tools 41

Similar servers

ketch

Bright Data MCP

Hyperbrowser MCP Server

SearXNG MCP Server

Crawl4AI

Scrapling MCP Server