Log inSubmit server
Wigolo logo

Wigolo

by KnockOutEZ

Local-first web search, browsing, crawling, extraction, caching, and research through a keyless MCP server.

InstallRepositoryWebsite
Full README Compare alternatives Embed widget Claim this listing

Install Wigolo

Pick your client, then add it in one click or paste the config.

STDIONode.js
~/Library/Application Support/Claude/claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "knockoutez-wigolo": {
      "command": "npx",
      "args": [
        "-y",
        "wigolo"
      ]
    }
  }
}

Settings → Developer → Edit Config, merge this into mcpServers, then fully quit and reopen Claude Desktop.

Install not yet sandbox-tested— we haven't run this command through our automated sandbox yet. Not a red flag; we're working through the catalog.

Overview

KnockOutEZ/wigolo gives AI agents a local web-intelligence layer for searching, fetching, crawling, extracting, caching, finding similar pages, and conducting research. It runs as an MCP server through npx, uses multiple search engines with local reranking, renders JavaScript-heavy pages, and stores its cache on the local machine. Core tools work without API keys; an LLM provider can optionally synthesize cited answers for research and agent workflows. Reach for it when an agent needs web evidence, structured page data, or multi-page collection without sending core activity to a hosted service.

Use cases

  • Search the web and return cited evidence for an agent
  • Fetch JavaScript-heavy pages and convert them to Markdown
  • Crawl a site while respecting robots.txt and rate limits
  • Extract tables or schema-based data from web pages
  • Reuse cached web results and detect page changes

Key features

  • Multi-engine search with local ML reranking
  • Headless-browser fallback for dynamic pages
  • BFS, DFS, sitemap, and map-only crawling
  • Structured extraction with custom JSON Schema
  • Persistent SQLite cache with semantic lookup
  • Citation-aware evidence and freshness signals

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked 1d ago

Recent health check succeeded.

28/28 checks healthy over the last 75d

Supply-chain signal

No high-severity advisories surfaced by our automated scan.

Critical 0High 0Medium 0Low 0

Scanned 8d ago via OSV.dev · wigolo (npm)

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

GitHub stars5.5kStargazers on the source repository.
npm downloads3.5kPackage downloads in the last 30 days.
Last commit1d agoMost recent push to the default branch.
Directory activity3 viewsConfig copies, upvotes, and views on AllMCPs.

How Wigolo works

What KnockOutEZ/wigolo does

KnockOutEZ/wigolo provides a single local interface for web research tasks. Its MCP tools cover web search, URL retrieval, multi-page crawling, structured extraction, cache queries, similarity discovery, and research-oriented gathering. The server is aimed at AI agents that need to find and inspect current web material rather than rely only on a model's existing knowledge.

Search combines results from multiple engines and applies local machine-learning reranking. Returned evidence can include excerpts, citation identifiers, source spans, scoring details, and freshness signals. This gives an agent material it can cite and evaluate instead of only receiving a list of links.

How it works

The server runs locally beside an agent or on the machine hosting self-managed agents. It can be registered as npx -y wigolo in an MCP client configuration, and it also supports REST, SDK, and embedded application usage. Core web operations are keyless, with data kept under ~/.wigolo/.

Fetching uses a tiered route that starts with ordinary HTTP and escalates to a headless browser for anti-bot challenges or JavaScript application shells. The fetch workflow can return cleaned Markdown, metadata, and links, and supports PDFs, selected sections, authenticated sessions, and browser actions such as clicking, typing, scrolling, and screenshots.

Crawling supports breadth-first, depth-first, sitemap, and map-only modes. It applies per-domain rate limits, respects robots.txt, and removes repeated boilerplate. A persistent SQLite cache supports keyword or hybrid semantic lookup, statistics, clearing, and change detection.

Setup and configuration

Installation requires Node 20 or newer and approximately 1.5 GB of free disk space on macOS, Linux, or Windows. Run npx wigolo init to install the local engine, browser engine, and on-device models, then perform a health check. npx wigolo init --agents=claude-code,cursor can also write configuration for named agents. Use npx wigolo doctor to check the installation later.

Initialization is unattended by default and supports --interactive, --wizard, and --no-warmup. The latter postpones downloads until first use. To remove the installation, run npx wigolo config --uninstall --yes.

The research, agent, and search answer format can use an LLM to produce synthesized cited responses. Gemini, Anthropic, OpenAI, Groq, Ollama, and OpenAI-compatible endpoints are supported through configuration; without an LLM, these workflows return a raw brief and evidence for the calling agent to assemble.

Tools and capabilities

  • search: Query multiple web search engines, rerank results locally, search in parallel with query arrays, filter by domain or time, match exact phrases, and request image results.
  • fetch: Retrieve and render pages, including JavaScript-heavy sites, PDFs, selected sections, and pages requiring actions or authenticated sessions.
  • crawl: Traverse sites using BFS, DFS, sitemap, or map-only strategies.
  • extract: Produce tables, metadata, JSON-LD, brand information, named schemas, or custom JSON Schema output.
  • cache: Search previously seen material with keyword or hybrid semantic matching, inspect statistics, clear entries, and detect changes.
  • find_similar, research, and agent: Discover related pages and support evidence-gathering workflows.

Limitations and notes

The project is marked as public beta. The main search, fetch, crawl, extract, cache, and similarity features do not require API keys. Finished synthesized answers for research-oriented workflows require an LLM provider, although a local Ollama or compatible endpoint can keep that workflow keyless.

Read the full README View source on GitHub

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about Wigolo

KnockOutEZ/wigolo is an MCP server that connects AI agents to the web through local, keyless tools for search, fetch, crawl, extract, cache, find-similar, research, and agent workflows. It uses multiple search engines, a browser engine, local reranking, and a persistent SQLite cache, with optional LLM synthesis for cited answers.

Run `npx wigolo init` with Node 20 or newer and about 1.5 GB of free disk space. To configure supported agents during setup, use `npx wigolo init --agents=claude-code,cursor`; other MCP clients can register `npx -y wigolo` in their own configuration.

No, its search, fetch, crawl, extract, cache, and find-similar tools work without an API key. The research, agent, and answer-format search workflows can use an LLM provider for synthesized responses, or return raw evidence for the calling agent to assemble.

Run `npx wigolo doctor` to check the local installation. The initial `npx wigolo init` command also downloads the browser engine and on-device models, runs a health check, and reports the readiness of each component.

KnockOutEZ/wigolo names Claude Code, Cursor, Codex, Gemini CLI, VS Code, Windsurf, and Zed as supported agents, and it can be registered with any MCP client. Its standard npx-based setup also fits clients that accept ordinary MCP server configurations.

Related MCP servers

All in Search & Data Extraction All alternatives
Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewWigolo AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/knockoutez-wigolo?style=directory)](https://allmcps.com/mcp/knockoutez-wigolo)
HTML Embed
<a href="https://allmcps.com/mcp/knockoutez-wigolo"><img src="https://allmcps.com/api/badge/knockoutez-wigolo?style=directory" alt="Wigolo on AllMCPs" /></a>