wigolo

Name: wigolo
Author: KnockOutEZ (yourtowhid)

Search & Research

KnockOutEZ (yourtowhid)

Open SourceFree

About

Local-first web intelligence for AI agents: search, fetch, crawl, extract, cache, find-similar, research, and autonomous gather loops — no API keys, no metered bill. Serves agents via MCP, REST, and SDKs.

Visit

Share this tool

Twitter LinkedIn Facebook

Our Verdict

Worth Trying

A genuinely free, local-first web layer for agents — keyless search with honest output, still young and AGPL.

wigolo aims to be the local-first answer to Firecrawl, Exa, and Tavily: 18 search engine adapters with on-device reranking, tiered fetch with headless-browser escalation, structured extraction, and a persistent local cache — all keyless and $0 per query. The honest-output philosophy (surfacing blocked_by_challenge failures and stale cache instead of hiding them) is a genuinely refreshing design choice.\n\nConsiderations: it needs ~1.5 GB local disk, the AGPL-3.0 license restricts commercial embedding, and it is maintained by a single person in public beta. For developers and self-hosters who want a private, zero-cost web layer for their agents, wigolo is worth trying.

Best for

•Developers and self-hosters who want a private, zero-cost web search/extraction layer for their AI agents, and value honest surfaced failures over opaque cloud APIs.

Consider alternatives if

•If you need a managed cloud API with an SLA, or a permissive license for commercial embedding, compare with Firecrawl (open source core) or the metered Exa/Tavily APIs.

Try wigolo

Supported Platforms

WindowsmacOSLinuxAPI

Available platforms include Windows, macOS, Linux, and API.

Key Features

Multi-engine web search (18 direct adapters) with rank fusion, ML reranking, and explainable per-result scores

Tiered fetch: auto-escalates from plain HTTP to a headless browser engine on anti-bot challenges or SPA shells

Multi-page crawl with BFS/DFS/sitemap modes, per-domain rate limits and robots.txt respect

Structured extraction: tables, JSON-LD, brand identity, named schemas, or custom JSON Schema

Persistent local cache with keyword/hybrid semantic query and change detection

find_similar via 3-way fusion of keyword + semantic + live web

research tool: decompose question → fan out sub-queries → synthesize cited report

autonomous agent gather loop with step log, time budget and output schema

diff + watch tools for page change detection with webhook delivery

Runs as MCP server, REST API (wigolo serve), CLI, interactive shell, and TypeScript/Python SDKs

Pricing

free

Free and open source (AGPL-3.0). Core tools are fully keyless; optional LLM synthesis can use a free Gemini key or local Ollama.

Use Cases

Giving coding agents (Claude Code, Cursor, Codex) a local-first web search and fetch layer without API keys

Building RAG pipelines with fresh web data, cached locally with change detection

Self-hosted agents, n8n workflows and any MCP client getting structured web extraction

Pros

Truly keyless: 18 search engine adapters with on-device reranking and embeddings — $0 per query

Honest output: surfaces stale cache, failed fetches, blocked_by_challenge failures instead of hiding them

Privacy by default: everything lives under ~/.wigolo/, nothing leaves the machine unless you opt in

Broad surface: MCP, REST, CLI, shell, TypeScript and Python SDKs

Local persistent cache makes re-queries instant and offline-capable

Cons

Requires ~1.5 GB local disk for browser engine and on-device models

AGPL-3.0 license — restrictive for commercial embedding in proprietary products

Public beta status; project is young (created April 2026)

Solo maintainer (personal account) — bus factor consideration

Search quality depends on public engines accessible from your network

Latest Update

August 2026: wigolo continues steady development with 10 tools (search, fetch, crawl, extract, cache, find_similar, research, agent, diff, watch), multi-agent wiring (claude-code, cursor, codex, gemini-cli, opencode, vs-code, windsurf, zed, antigravity), REST API, and TypeScript/Python SDKs.

Related Search & Research Tools

Perplexity

Perplexity AI

AI search engine providing real-time answers with cited sources

Phind

AI search engine for developers that provides direct technical answers with code examples, more efficient than traditional search

Metaso

Metasota

A leading Chinese AI search engine by Metasota. Ad-free with direct answers, it auto-generates outlines, mind maps and cited sources, supports academic search and deep research modes, and offers an open API

Felo

Sparticle

A multilingual AI search engine by Japan's Sparticle. It searches global sources across languages with cited answers, and also generates presentations, mind maps and documents, with a generous free tier