verify-api MCP server
io.github.drain54/verify-apidrain54streamable-httpv1.0.0Scanned
AI claim verification MCP endpoint with x402 payments and stealth web fetch
8 tools: 5 change things and 3 only read.
- Reads private data · 0
- Reads text others write · 3
- Can send data out · 5
- 1 finding
Sign-in required, no OAuth metadata. Remote posture
Anlyon alternative
5 of these tools change something. Run those as governed actions.
Declare what an action changes, and Anlyon puts it under a shared limit, binds every approval to the exact request, and grades the outcome. It works on any HTTPS API you can describe as an action.
Anlyon Radar does not run, proxy or govern this MCP server. A tool your own code calls directly is not routed through Anlyon and is not governed by it.
Import an OpenAPI spec →What the verify-api MCP server can do
The verify-api MCP server is published by the GitHub account drain54. It has 8 tools. It can create or change data (5 tools: verify_ai_claim, check_endpoint_health, solve_captcha, search_web and 1 more), but cannot delete data. 3 tools only read.
3 tools bring in text other people can write (read_web_page, search_web and fetch_stealth_web); 5 tools can send data out of your system (check_endpoint_health, solve_captcha, read_web_page and 2 more).
The fixed checks matched 1 finding (annotation-conflict). Each one comes with the text it matched.
Last scanned Nov 5, 2026, version 1.0.0.
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.
8 tools
Changes things 5Claude Code: ask
check_endpoint_healthProbe and verify the reachability, HTTP status code, latency, and operational health of an AI API endpoint or web service.Can 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
- 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 Probe and verify the reachability, HTTP status code, latency, and operational health of an AI API endpoint or web service. Follows redirects with a 15-second timeout. Requires x402 micropayment (0.01 USDC on Base). When to use: Check if a specific API URL or model endpoint is online, responding, or returning 5xx/402 errors. When NOT to use: Do NOT use for general domain WHOIS or DNS record lookups. Parameters: - `url` (string, required): Target HTTP/HTTPS endpoint URL to probe (8-1000 chars), e.g. 'https://api.openai.com/v1/models'.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
extract_json_from_webExtract structured JSON data matching a custom schema directly from any webpage.Can 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
- 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 Extract structured JSON data matching a custom schema directly from any webpage. Extracts clean content and processes schema mapping via fast LLM parsing. Requires x402 micropayment (0.015 USDC on Base). When to use: Scraping structured data (product specs, prices, jobs, articles) into clean JSON. When NOT to use: Do NOT use for general open-ended chat without a defined schema. Parameters: - `url` (string, required): Target webpage URL. - `schema` (object, required): JSON object describing fields or schema to extract. - `instructions` (string, optional): Specific guidance for extraction.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
search_webExecute live web searches using multi-engine chain (TinyFish, DuckDuckGo, Jina) without monthly API subscriptions.Reads text others write1 finding
Why it is in this group
- The name "search_web" says "search": a word for reading.
- The description's first verb is "execute": a word for a change.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
What it can reach
- untrusted-inputThe name "search_web" says "search" and "web": it reads web pages, which anyone can write.
- untrusted-inputThe description's first sentence says "execute" and "web": it reads web pages, which anyone can write.
Findings
- Warn
The server marks "search_web" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true}; The name "search_web" says "search": a word for reading.
Description, as the server wrote it Execute live web searches using multi-engine chain (TinyFish, DuckDuckGo, Jina) without monthly API subscriptions. Returns fresh source URLs, titles, and snippets. Requires x402 micropayment (0.005 USDC on Base). When to use: Real-time web browsing and information retrieval for AI agents. When NOT to use: Do NOT use for deep recursive crawling of entire sites. Parameters: - `query` (string, required): Search query (3-300 chars). - `limit` (integer, optional, default 5): Maximum number of search results to return (1-10).
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
solve_captchaSolve web CAPTCHA challenges (Turnstile, hCaptcha, reCAPTCHA v2, Arkose, Cloudflare) and return a valid solution token.Can 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.
What it can reach
- 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 Solve web CAPTCHA challenges (Turnstile, hCaptcha, reCAPTCHA v2, Arkose, Cloudflare) and return a valid solution token. Automatically retries once on failure without double charging. Requires x402 payment on Base. When to use: Use when an agent encounters a bot wall or CAPTCHA challenge during automated web workflows. When NOT to use: Do NOT use for non-CAPTCHA auth, 2FA/OTP codes, or general login forms. Parameters: - `type` (string, required): CAPTCHA type ('turnstile', 'hcaptcha', 'recaptcha', 'arkose', 'cloudflare'). - `sitekey` (string, required): Public sitekey extracted from the target page DOM. - `url` (string, required): Full target webpage URL hosting the challenge.
Annotations the server declares
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
verify_ai_claimVerify whether an AI model, API provider, pricing claim, or infrastructure assertion is true today using live web evidence.No 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.
Description, as the server wrote it Verify whether an AI model, API provider, pricing claim, or infrastructure assertion is true today using live web evidence. Returns deterministic verdicts (TRUE, FALSE, PARTIALLY_TRUE, UNREACHABLE, UNVERIFIED) with cited sources and confidence scores. Requires x402 micropayment (0.01 USDC on Base). When to use: Fact-checking an AI provider's claims, pricing, or model availability. When NOT to use: Do NOT use for general open-ended web search, coding assistance, or non-AI claim verification. Parameters: - `query` (string, required): The exact claim or assertion to verify (5-500 chars), e.g. 'Is GLM-5.3 Flash free on ZenMux?'. - `depth` (string, optional, default 'standard'): Verification depth. 'standard' executes fast single-pass search (0.01 USDC); 'deep' conducts multi-source cross-examination (0.03 USDC).
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
Only reads 3Show 3 read-only tools
fetch_stealth_webFetch webpage HTML with modern browser TLS and header impersonation (Sec-Ch-Ua, realistic headers) to bypass bot protection and detect anti-bot challenges.Reads text others writeCan send data outNo findings
Why it is in this group
- The name "fetch_stealth_web" says "fetch": a word for reading.
What it can reach
- untrusted-inputThe name "fetch_stealth_web" says "fetch" and "web": it reads web pages, which anyone can write.
- untrusted-inputThe description's first sentence says "fetch" and "webpage": it reads web pages, which anyone can write.
- untrusted-inputThe input takes a URL (input.url), and the tool reads it: it reads web pages, which anyone can write.
- outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
Claude Code: ask, because it is tagged outbound.
Description, as the server wrote it Fetch webpage HTML with modern browser TLS and header impersonation (Sec-Ch-Ua, realistic headers) to bypass bot protection and detect anti-bot challenges. Requires x402 micropayment (0.01 USDC on Base). When to use: Fetching websites that block standard cURL or basic HTTP libraries with 403 Forbidden. When NOT to use: Do NOT use for downloading giant binary files (videos, zip archives). Parameters: - `url` (string, required): Target webpage URL. - `custom_headers` (object, optional): Additional HTTP headers to pass along.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
get_captcha_pricingRetrieve current x402 pricing per 1,000 CAPTCHA solves across all supported types (Turnstile, hCaptcha, reCAPTCHA v2, Arkose, Cloudflare).No findings
Why it is in this group
- The name "get_captcha_pricing" says "get": a word for reading.
Description, as the server wrote it Retrieve current x402 pricing per 1,000 CAPTCHA solves across all supported types (Turnstile, hCaptcha, reCAPTCHA v2, Arkose, Cloudflare). Free endpoint with zero parameters. When to use: Check current rates and atomic USDC requirements before calling solve_captcha. When NOT to use: Do NOT use to submit or solve CAPTCHA challenges.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
read_web_pageExtract clean, readable Markdown and metadata from any public webpage for LLM ingestion, stripping ads, popups, and navigational clutter.Reads text others writeCan send data outNo findings
Why it is in this group
- The name "read_web_page" says "read": a word for reading.
What it can reach
- untrusted-inputThe name "read_web_page" says "read" and "web": it reads web pages, which anyone can write.
- untrusted-inputThe description's first sentence says "webpage": it reads web pages, which anyone can write.
- untrusted-inputThe input takes a URL (input.url), and the tool reads it: it reads web pages, which anyone can write.
- outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
Claude Code: ask, because it is tagged outbound.
Description, as the server wrote it Extract clean, readable Markdown and metadata from any public webpage for LLM ingestion, stripping ads, popups, and navigational clutter. Returns clean markdown, title, description, character count, and estimated tokens. Requires x402 micropayment (0.005 USDC on Base). When to use: Ingesting articles, blog posts, documentation, or news pages into LLM context. When NOT to use: Do NOT use for raw binary files (PDF/images), authenticated pages behind a login, or single-page apps that require heavy JavaScript rendering. Parameters: - `url` (string, required): Full target webpage URL (e.g. 'https://news.ycombinator.com'). - `include_links` (boolean, optional, default true): Whether to preserve markdown hyperlinks. - `include_images` (boolean, optional, default false): Whether to preserve image markdown links.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: 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 write3
- Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
- Can send data out5
- 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.
Looked at Nov 20, 2026, 19:04 UTC, unchanged since Nov 20, 2026. No credentials were sent and no tool was called. What is looked at.
- Reachable
- yes
- HTTP status of an anonymous initialize
- 530
- tools/list served without credentials
- no
- Tools that change things, listed without credentials
- 0
- Protocol version
- not observed
- 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
- yes
- Dynamic client registration
- not observed
- Client ID metadata documents
- not observed
The probe stopped: initialize answered HTTP 530.
Posture changes
- Nov 20, 2026
- HTTP status of an anonymous initialize: 200 to 530
- tools/list served without credentials: yes to no
- Tools that change things, listed without credentials: 5 to 0
- Protocol version: 2025-06-18 to none
Findings
Matches of fixed checks, with the text they matched.
- Warnsearch_web
The server marks "search_web" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true}; The name "search_web" says "search": a word for reading.
History
Changes to its tool definitions and to its remote endpoint, newest first.
PostureRemote posture changed. HTTP status of an anonymous initialize: 200 -> 530; tools/list served without credentials: yes -> no; Tools that change things, listed without credentials: 5 -> 0 (and 1 more).View diff
- HTTP status of an anonymous initialize: 200 to 530
- tools/list served without credentials: yes to no
- Tools that change things, listed without credentials: 5 to 0
- Protocol version: 2025-06-18 to none
Install
Pinned to version 1.0.0. Values in angle brackets are yours to fill in.
claude mcp add --transport http verify-api https://verify.drain54.my.idclaude mcp add --transport http verify-api https://verify.drain54.my.idRun 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.
{
"permissions": {
"allow": [
"mcp__verify-api__get_captcha_pricing"
],
"ask": [
"mcp__verify-api__verify_ai_claim",
"mcp__verify-api__check_endpoint_health",
"mcp__verify-api__solve_captcha",
"mcp__verify-api__read_web_page",
"mcp__verify-api__search_web",
"mcp__verify-api__extract_json_from_web",
"mcp__verify-api__fetch_stealth_web"
],
"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 verify-api. Review it before you use it.
Capability tags behind 6 rules
- mcp__verify-api__check_endpoint_health outbound
- mcp__verify-api__solve_captcha outbound
- mcp__verify-api__read_web_page untrusted-input, outbound
- mcp__verify-api__search_web untrusted-input
- mcp__verify-api__extract_json_from_web outbound
- mcp__verify-api__fetch_stealth_web 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/io.github.drain54/verify-api/verify?hash=03a76a9d5ce0ae6ce0105757126defc5c2120a810ccdd74c71c7e576f1834e0e" | jq -e '.changed == false'Pinned hash 03a76a9d5ce0, from Nov 5, 2026, 17:18 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/io.github.drain54/verify-api/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/io.github.drain54/verify-api)Source: MCP Registry · repository. Synced Oct 21, 2026, 18:33 UTC.
verify-api MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the verify-api MCP server safe?
- No scan can promise that. What its definitions show: 8 tools, 5 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 verify-api MCP server delete data?
- No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.
Can the verify-api MCP server do things that cannot be undone?
- No tool is classed as irreversible: none of them sends, pays, publishes or merges, going by their definitions.
What tools does the verify-api MCP server have?
- 8 tools: 5 write, 3 read. verify_ai_claim, check_endpoint_health, solve_captcha, search_web, extract_json_from_web, get_captcha_pricing, read_web_page and fetch_stealth_web.
Does the verify-api MCP server need sign-in or OAuth?
- Its remote endpoint did not answer the last anonymous look.
Does the verify-api MCP server send data to outside services?
- 5 tools can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): check_endpoint_health, solve_captcha, read_web_page, extract_json_from_web and fetch_stealth_web.
Has the verify-api MCP server changed its tools?
- Not since Anlyon Radar started tracking it. Last scanned Nov 5, 2026.