Axongate

io.github.sauliusbeconisPublished by GitHub account sauliusbeconisv2.1.0Updated Oct 30, 2026

Safe web read, search, cited answers, fact-checks and token checks for agents, paid via x402.

Overview

AI-generated overview

AxonGate gives an assistant paid web reading, search, cited answers, fact-checks and crypto token safety checks, billed per call in USDC on Base.

What you can do
  • Read a URL into clean Markdown with a prompt-injection scan, or read up to 20 URLs in one batch
  • Search the live web, optionally reading the top results
  • Scan untrusted text for prompt-injection attempts
  • Produce answers and claim verdicts backed by verbatim-checked citations
  • Check a crypto token for honeypot, tax, owner-power and liquidity risks
When to use it
Useful when an assistant needs current web content, sourced answers or claim verification, and when untrusted pages or text must be screened for prompt injection. Also relevant for quick automated risk signals on a crypto token.
Requirements
Runs locally over stdio via the npm package axongate-mcp (Node.js with npx). Needs a dedicated wallet private key in AXONGATE_PRIVATE_KEY holding a few dollars of USDC on Base; no ETH is required. AXONGATE_MAX_SPEND_USDC sets the per-process spend cap (default 0.50). No account or API key.
Before you install
The server holds a wallet private key in AXONGATE_PRIVATE_KEY and spends real USDC on every call, so use a dedicated low-balance wallet and keep the spend cap low. Token check verdicts are automated risk signals from third-party sources, not financial advice.

Installation

In SourceWeft

  1. Open Axongate in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

axongate-mcp

MCP server for AxonGate: safe web access and token checks for AI agents, paid per call in USDC on Base over x402. No account or API key; failed calls are never charged.

ToolWhat it doesPrice
axongate_readURL → clean Markdown + prompt-injection scan$0.002
axongate_read_batchUp to 20 URLs in one call, each scanned$0.03
axongate_searchLive web search, optionally reading the top results$0.004
axongate_scanPrompt-injection scan of untrusted text$0.002
axongate_answerAnswer with verbatim-checked citations$0.015
axongate_verifyClaim verdict with verbatim-checked evidence$0.02
axongate_token_checkCrypto token safety verdict (honeypot, taxes, owner powers, liquidity)$0.01

Setup

Use a dedicated wallet holding a few dollars of USDC on Base. No ETH is needed: the facilitator pays gas. AXONGATE_MAX_SPEND_USDC caps total spend per server process (default 0.50).

Claude Code:

bash
claude mcp add axongate -e AXONGATE_PRIVATE_KEY=0x... -e AXONGATE_MAX_SPEND_USDC=1.00 -- npx -y axongate-mcp

Claude Desktop, Cursor and other MCP clients:

json
{  "mcpServers": {    "axongate": {      "command": "npx",      "args": ["-y", "axongate-mcp"],      "env": { "AXONGATE_PRIVATE_KEY": "0x...", "AXONGATE_MAX_SPEND_USDC": "1.00" }    }  }}

Token Check data is powered by GoPlus Security, with honeypot.is simulations where available. Its verdicts are automated risk signals, not financial advice.

Source: mcp/README.md at commit bab4f07

Tools

0
SourceWeft started this package once in an isolated sandbox, without credentials, to list its tools. That failed: the server did not start.

Version history

1
  1. v2.1.0LatestOct 29, 2026

Score

50% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts