mcpField
V
verantis-mcp

verantis-mcp

Remote
github.com/m9labs-railscopeFinance4Updated 8w ago
Repository

About verantis-mcp

verantis-mcp lets MCP-capable agents inspect machine-payment services and recipient wallets before payment. It reads Verantis's public directory of x402 and MPP services, including verification provenance and reputation information.

Key features

  1. 01Search the x402 and MPP directory
  2. 02Filter services by price, chain, and reputation
  3. 03Retrieve service records and verification provenance
  4. 04Check wallet reputation and buyer retention

Use cases

  1. 01Find verified machine-payment services
  2. 02Review a service before an agent pays
  3. 03Check a payment recipient wallet

MCP support

Exposes read-only directory search, service lookup, wallet checks, and directory statistics as MCP tools. It can be used locally over stdio or through the hosted Streamable HTTP endpoint.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttps://api.verantis.ai/mcp
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.

Packages
pypiverantis-mcp0.2.1

Tools

4
ParameterTypeRequiredDescription
chainstringNosettlement network, e.g. base, solana, polygon, tempo
limitintegerNomaximum number of results to return (default 10)
querystringYeswhat you need, e.g. 'twitter data', 'web search', 'rpc ethereum'
protocolenumNorestrict to one payment protocol: x402 or mppx402 | mpp
min_scoreintegerNominimum trust score 0-100 — set high (e.g. 85) for money-moving tasks
proven_onlybooleanNoonly services that have EARNED reputation over time (excludes 'new' services and 'watch' services whose on-chain payments are concentrated / not yet broadly distributed). Reputation is earned via consistent history + on-chain buyer retention; a brand-new well-behaved service is verified but new, not top-trust.
max_price_usdnumberNomax price per call in USD
require_freshbooleanNolive re-probe stale results before returning — use before paying real money; adds ~1s per stale service
verified_onlybooleanNoonly services that passed the latest live probe
max_latency_msintegerNoreject services slower than this (ms)
exclude_concentratedbooleanNoexclude services from operators with many listings under one registrable owner (a neutral concentration signal, not a fraud claim)

Details

From the project README

Verantis MCP

The verified directory for machine payments — as an MCP tool.

Install in VS Code PyPI MCP Registry

Give any MCP-capable agent (Claude, Cursor, agent frameworks) the ability to check what it's about to pay before it pays. Verantis continuously probes machine-payable services across x402 and MPP, measures on-chain reputation, and publishes a recomputable record. This server is a thin client of the public Verantis API — standard library only, no dataset needed.

Learn more: verantis.ai · methodology: verantis.ai/methodology

Tools

ToolWhat it does
find_paid_serviceSearch the verified directory (x402 / MPP). Filter by verified-only, price, chain, protocol, category, reputation.
get_serviceFull record + verification provenance for one domain.
check_walletPre-payment guard — pass the recipient wallet a service asked you to pay; get its earned reputation tier, on-chain buyer retention, and which services it fronts.
directory_statsIndex-level stats: services, verification breakdown, freshness.

What it can and can't do

  • Can: read the public directory; check a service or wallet before your agent pays.
  • Can't: move money, see your keys, or read your files.
  • Only ever calls api.verantis.ai — nothing else.
  • Writes nothing (hosted endpoint) / only a random anonymous install id (local pip/uvx).
  • No account, no API key. MIT-licensed. Read-only.
  • How we measure: verantis.ai/methodology

Install & connect

Two ways to connect — pick one.

Local (stdio) — via uvx, no install step

Claude Desktop (claude_desktop_config.json) or any MCP client:

{
  "mcpServers": {
    "verantis": {
      "command": "uvx",
      "args": ["verantis-mcp"]
    }
  }
}

Prefer pip? pip install verantis-mcp, then use "command": "verantis-mcp" with no args.

Cursor (~/.cursor/mcp.json): same block under mcpServers.

Hosted (no install) — connect by URL

VS Code / Copilot: one-click add ▸ — or add it manually below.

For any client that supports remote / streamable-HTTP MCP servers, point it straight at the hosted endpoint — nothing to install, no Python:

{
  "mcpServers": {
    "verantis": { "url": "https://api.verantis.ai/mcp" }
  }
}

Clients that only speak local stdio can bridge with npx mcp-remote https://api.verantis.ai/mcp.

Restart your client, and the four tools appear.

Configuration

Env varDefaultPurpose
VERANTIS_API_BASEhttps://api.verantis.aiPoint the client at a different Verantis API.

Notes

Informational only; not financial, investment, or legal advice. Classifications describe what live probes and on-chain reads observed at a point in time and can change. Reputation is earned from real on-chain settlement, not assumed. Verantis has no rail of its own, no token, and no stake in any listed service.

MIT licensed.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.