Skip to content

foura.ai MCP server

ai.foura/mcpfoura.aistreamable-httpv0.8.4Scanned

Reliable web access for AI agents: smart HTTP, rotating proxies, and full-browser rendering.

4 tools: 1 do something that can't be undone and 3 change things.

Tools: 0 delete, 1 irreversible, 3 write, 0 read.
  • Reads private data · 0
  • Reads text others write · 4
  • Can send data out · 4
  • 1 finding

Tools listed without sign-in. Remote posture

With Anlyon

1 tool here does something that can't be undone. Govern it.

  • Anlyon lists the tools and adopts none.
  • Each tool you adopt gets approvals and receipts.
  • Your agent never holds the key.
Govern this server →

A tool your agent calls directly is not routed through Anlyon and is not governed by it. Anlyon Radar only lists and labels, and these labels are the same either way.

What the foura.ai MCP server can do

The foura.ai MCP server is published by foura.ai, a domain the MCP Registry verified. It has 4 tools. It can do things that cannot be undone, such as sending, paying or publishing (foura_single) and create or change data (3 tools: foura_proxy, foura_browser and foura_auto), but cannot delete data. None of its tools only reads.

4 tools bring in text other people can write (foura_single, foura_proxy, foura_browser and 1 more); 4 tools can send data out of your system (foura_single, foura_proxy, foura_browser and 1 more).

The fixed checks matched 1 finding (annotation-conflict). Each one comes with the text it matched.

Its remote endpoint lists its tools without sign-in.

Last scanned Nov 18, 2026, version 0.8.4.

Tools

Grouped by what each tool does, most consequential first. Open one for its reasons and full description; descriptions are the publisher's own words.

4 tools

Can't be undone 1Claude Code: ask
  • foura_singleSend one HTTP request and return the response.Reads text others writeCan send data out1 finding

    Why it is in this group

    • The description's first verb is "send": a word for sending, and what is sent cannot be taken back.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • outboundThe description's first verb is "send": it sends or publishes something to other people.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.

    Findings

    • Warn

      The server marks "foura_single" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"openWorldHint":true}; The description's first verb is "send": a word for sending, and what is sent cannot be taken back.

      What this check looks for
    Description, as the server wrote it
    Send one HTTP request and return the response. Use it for static pages, JSON APIs, and server-rendered HTML, with your own headers, body and cookies. Set browser, os, or version when a target refuses the default Chrome. Pass a proxy id returned by foura_proxy to leave from that same exit. validate decides what counts as success, so a response it rejects comes back as a failure instead of as content; the timeouts nest, with timeout_ms bounding the whole request and the connect, accept and server_response ones bounding stages inside it. Switch to foura_proxy when the response is blocked or the exit country matters, and to foura_browser when the page needs JavaScript. One FourA API key authenticates every call, the result reports the credits it spent, and a refusal by your own plan arrives as a plan_limit_ code with retryAfter rather than as a block by the target.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: true
Changes things 3Claude Code: ask
  • foura_autoGive it a public URL and get the content back.Reads text others writeCan send data outNo findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
    Description, as the server wrote it
    Give it a public URL and get the content back. Use it when only the content matters and not how it is fetched: it escalates from a direct request to a rotating proxy to a full browser only as far as the target forces, and returns the session that worked so the next call can replay it. On protected targets, or whenever HTTP 200 may still be a challenge or an incomplete page, pass validate.data.accept with text unique to the real content; auto makes bounded attempts and returns either validated content or a failure, and cannot guarantee a match. It owns its own retry settings, so there is no maxTries here, and timeout_ms is the budget for every attempt together rather than for one. What it cannot do is the reason to reach for another tool: a strict exit country and a pinned exit live on foura_proxy (auto can only avoid exits, through ignoreProxies, or escalate to a premium exit with exitClass); choosing which browser is presented to the target lives on foura_single and foura_proxy; a scripted browser session lives on foura_browser. A request that names any of those belongs there rather than here. One FourA API key authenticates every call, the result reports the credits it spent, which is the sum of the attempts it made, and a refusal by your own plan arrives as a plan_limit_ code with retryAfter rather than as a block by the target.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: true
  • foura_browserLoad a public URL in a full browser session.Reads text others writeCan send data outNo findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
    Description, as the server wrote it
    Load a public URL in a full browser session. JavaScript runs, the DOM renders, and cookies come back with the response. Use it for single-page apps, lazy-loaded content, or supported browser challenges. You can set your own headers, cookies and userAgent, and checkStatus or checkText turn a navigation that rendered the wrong page into a failure instead of returning it as content. For a protected page, call foura_proxy first and pass its returned proxy ID here to reuse that same exit, instead of starting a new selection here; with no proxy the session leaves from one fixed address. Set unblocker:false when you want the page exactly as it loads, challenge included. It is the most expensive of the four tools and the slowest, so prefer foura_single or foura_proxy whenever the content is already in the HTML. One FourA API key authenticates every call, the result reports the credits it spent, and a refusal by your own plan arrives as a plan_limit_ code, including plan_limit_browser_daily when the daily browser allowance is spent.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: true
  • foura_proxyRoute an HTTP request through rotating proxies, retrying on another exit until one delivers.Reads text others writeCan send data outNo findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • outboundThe input takes a URL (input.request.url): a request to a URL can carry data to whoever runs it.
    Description, as the server wrote it
    Route an HTTP request through rotating proxies, retrying on another exit until one delivers. Use it when foura_single is blocked, and whenever the exit itself matters: this is the only tool that takes a strict exit-country allowlist (exitCountries, which never falls back to another country), presents a named browser family per attempt, excludes exits you already know are dead (ignoreProxies), and may escalate to a premium exit (exitClass, an allowance rather than an instruction). maxTries bounds how many exits are tried and timeout_ms bounds the whole rotation, so a short timeout can end it before maxTries is reached; protected targets often need 25 to 30. validate decides what counts as delivered, and an attempt it rejects is retried on the next exit. The response returns the id of the exit that succeeded: pass that returned id to foura_single.proxy or foura_browser.proxy to send the next request from the same exit. A failed rotation returns attemptReport, which separates exits that never answered from exits a bot check refused from pages your own rule threw away. Use foura_browser when the page needs JavaScript. Rotation costs several times a single request. One FourA API key authenticates every call, the result reports the credits it spent, and a refusal by your own plan arrives as a plan_limit_ code with retryAfter rather than as a block by the target.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: true

This is not a safety rating, and no finding is not a safety guarantee. The groups and findings come from fixed checks of the tool definitions the server returned on the date shown, not from its code or what it does when called. How the checks work.

Exfil surface

Which tools read private data, which read text someone else can write, and which can send data out. One of each is a path a prompt injection can use.

No private-data, untrusted-input, outbound path in this server alone: no tool is tagged private-data.

Reads private data0
Reads files, mail, messages, databases, secrets, calendars or repositories.
Reads text others write4
Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
Can send data out4
Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.

Other servers in the same agent add their own tools to these paths. Audit your whole setup to see them together.

Remote posture

What anyone on the internet sees at the remote endpoint: an anonymous initialize and tools/list, and the public OAuth metadata.

Tools listed without sign-in

Looked at Nov 20, 2026, 18:25 UTC, unchanged since Oct 22, 2026. No credentials were sent and no tool was called. What is looked at.

Reachable
yes
HTTP status of an anonymous initialize
200
tools/list served without credentials
yes
Tools that change things, listed without credentials
4
Protocol version
2025-06-18
WWW-Authenticate on 401
not observed
resource_metadata in WWW-Authenticate
not observed
RFC 9728 protected resource metadata
not observed
Authorization servers
not observed
Legacy HTTP+SSE transport
no
Dynamic client registration
not observed
Client ID metadata documents
not observed
  • HighThe whole server

    The endpoint served tools/list without credentials, and the list has 4 tools that change things (1 irreversible, 3 write). Whether a call also works without credentials was not tested: tools are never called.

    Matched text

    https://mcp.foura.ai/mcp: tools/list returned 4 tools to an anonymous client, including foura_single (irreversible), foura_proxy (write), foura_browser (write), foura_auto (write)

    What this check looks for

Findings

Matches of fixed checks, with the text they matched.

1 warn
  • The server marks "foura_single" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"openWorldHint":true}; The description's first verb is "send": a word for sending, and what is sent cannot be taken back.

    What this check looks for

History

Changes to its tool definitions and to its remote endpoint, newest first.

No change seen yet. Its tools are rescanned every day, and a remote endpoint is looked at once a week.

Install

Pinned to version 0.8.4. Values in angle brackets are yours to fill in.

claude mcp add --transport http foura https://mcp.foura.ai/mcp
claude mcp add --transport http foura https://mcp.foura.ai/mcp

Run in a terminal.

Claude Code permissions

Allow the reads, ask before a change or a way out, deny the deletes. Paste into .claude/settings.json.

.claude/settings.json
{
  "permissions": {
    "allow": [],
    "ask": [
      "mcp__foura__foura_single",
      "mcp__foura__foura_proxy",
      "mcp__foura__foura_browser",
      "mcp__foura__foura_auto"
    ],
    "deny": []
  }
}

Tools that only read are allowed, tools that change things ask first, tools that delete or cancel are denied. A read that can send data out (it takes a URL, say) asks first too. It assumes the server is installed as foura. Review it before you use it.

Capability tags behind 4 rules
  • mcp__foura__foura_single untrusted-input, outbound
  • mcp__foura__foura_proxy untrusted-input, outbound
  • mcp__foura__foura_browser untrusted-input, outbound
  • mcp__foura__foura_auto untrusted-input, outbound

Pin

Keep the decision: fail CI when the tools change, follow every change, or show a badge.

Pin in CI

Fail a build when the tools you reviewed are not the tools the server serves today.

curl -s "https://radar.anlyon.com/api/v1/servers/ai.foura/mcp/verify?hash=76a36f7780d7df0694fa9fa2a1dcddf62c4c8891cc9d59288fae1c3b87f0111b" | jq -e '.changed == false'

Pinned hash 76a36f7780d7, from Nov 18, 2026, 15:18 UTC.

Watch

Get told when its tools change: a new tool, a new description, a new class.

https://radar.anlyon.com/servers/ai.foura/mcp/changes.xml

An Atom feed. Works in any feed reader or Slack's RSS app.

Badge

Anlyon Radar badge for ai.foura/mcp
[![Anlyon Radar](https://radar.anlyon.com/badge/ai.foura/mcp.svg)](https://radar.anlyon.com/servers/ai.foura/mcp)

Source: MCP Registry · repository. Synced Nov 18, 2026, 15:17 UTC.

foura.ai MCP server: questions and answers

Answered from the scan above, not from the publisher's description.

Is the foura.ai MCP server safe?

No scan can promise that. What its definitions show: 4 tools, 4 of which can change, send or delete something, and 1 finding from fixed checks. Allow the reads, ask before each change, and deny the deletes; the permissions block on this page does that.

Can the foura.ai MCP server delete data?

No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.

Can the foura.ai MCP server do things that cannot be undone?

Yes. 1 tool can send, pay, publish or merge: foura_single.

What tools does the foura.ai MCP server have?

4 tools: 1 irreversible, 3 write. foura_single, foura_proxy, foura_browser and foura_auto.

Does the foura.ai MCP server need sign-in or OAuth?

Its remote endpoint listed its tools to an anonymous client, without sign-in.

Does the foura.ai MCP server send data to outside services?

4 tools can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): foura_single, foura_proxy, foura_browser and foura_auto.

Has the foura.ai MCP server changed its tools?

Not since Anlyon Radar started tracking it. Last scanned Nov 18, 2026.