seo-tools-mcp-aparser

MCP server that bridges a self-hosted A-Parser instance to Claude Code and any MCP client — SERP, suggests and hundreds of parsers via A-Parser's HTTP API. Read-only, strict JSON output. Part of seo-tools-mcp (SEO servers for the Google/Yandex market).

You need your own running A-Parser instance (licence + server). This bridge drives it; it does not host or proxy A-Parser for you.

Install

claude mcp add aparser --scope user -- npx -y seo-tools-mcp-aparser

Self-contained package — the shared code is bundled in, nothing else to install. For any other MCP client (Claude Desktop, Cursor…), add one block to mcpServers:

{
  "mcpServers": {
    "aparser": {
      "command": "npx",
      "args": ["-y", "seo-tools-mcp-aparser"],
      "env": { "APARSER_URL": "http://IP:9091/API", "APARSER_PASSWORD": "..." }
    }
  }
}

Then set credentials right in the chat: aparser_auth_status → aparser_set_credentials. In A-Parser: Settings → API, enable the API server, note the port (usually 9091) and password. Check readiness with aparser_ping and aparser_status. Full docs and all servers: https://github.com/antohins/seo-tools-mcp

Tools

• aparser_ping — connectivity + password check
• aparser_status — readiness verdict: version, installed parsers, queue, live proxies (ready:false if 0 live proxies)
• aparser_proxies — live proxies on the instance (checkers filters to specific proxy packs); credentials never shown
• aparser_parsers — parsers installed on the instance
• aparser_parser_fields — result fields a parser can return (flat + arrays)
• aparser_get_preset — read a config preset's options (useproxy, domain, hl/gl, proxy checkers…)
• aparser_serp_google — Google organic SERP (SE::Google); proxies on by default + preflight
• aparser_serp_yandex — Yandex organic SERP (SE::Yandex); region via lr
• aparser_suggest — Google/Yandex search suggestions
• aparser_request — universal single request to any parser (oneRequest)
• aparser_bulk_request — bulk request: one parser, many queries, N threads (bulkRequest)

Proxies. A-Parser needs proxies + a running Proxy Checker to scrape Google/Yandex without bans. Proxy packs (proxy checkers) are configured once in the A-Parser GUI; this bridge reads, verifies and selects them (checkers), but does not create or load them. v1 is synchronous and read-only — the task queue and large async exports are a v2 candidate.

🛰 Maintained by PBN Workers — search-visibility infrastructure: semantic cores, PBN & satellites, SEO automation. We use these tools in production. Need steady organic traffic? Get in touch.

MIT © antohins