awesome-repositories.com
Blog
MCP
awesome-repositories.com

Discover the best open-source repositories with AI-powered search.

ExploreCurated searchesOpen-source alternativesSelf-hosted softwareBlogSitemap
ProjectMCP serverAboutHow we rankPress
LegalPrivacyTerms
© 2026 Bringes Technology SRL·VAT RO45896025·hello@awesome-repositories.com
KnockOutEZ avatar

KnockOutEZ/wigolo

0
View on GitHub↗
wigolo.app
↗

Wigolo

Wigolo is a local-first web research toolkit for AI agents. It runs as a Model Context Protocol server that gives coding agents web search, page fetching, crawling, and caching without API keys or cloud accounts. Search queries many public engines in parallel, merges results with rank fusion, and reranks them with on-device models, attaching an explainable relevance score to each result.

What sets it apart is the evidence it returns. Results carry verbatim excerpts pinned to positions in the source page, quotable citation IDs, and inspectable scores. Weak results are flagged as junk, and stale cache, failed engines, truncation, and bot-protected pages are reported as labeled failures rather than passed off as real content. An autonomous research mode plans, searches, fetches, and synthesizes cited answers within a time budget, using a configurable hosted or local language model, or returns a raw brief and evidence when none is set.

The same tools are available through MCP, a REST API with an OpenAPI contract, a command line and interactive shell, and typed TypeScript and Python SDKs. Page fetching escalates from plain HTTP to a headless browser when needed, and can click, type, scroll, and reuse signed-in sessions. Output is clean markdown or structured JSON, with fields checked against the source text. Site crawling honors crawl rules and per-domain rate limits, and page change tracking can notify a webhook. Everything retrieved is cached in a local keyword and semantic index on the user's machine, so repeat lookups work offline. Nothing is sent to third parties unless an external language model is explicitly enabled.

Wigolo can be extended with search engine plugins, a proxy and custom certificate authority, and skill packs and setup commands for named coding agents. It installs through several package registries and Docker images, and includes health checks that name any broken component and its fix. Core search, fetch, crawl, extract, and cache features keep working if the optional models or browser fail to install.

Features

  • Keyless Local Web Search - A locally run Model Context Protocol server that gives AI coding agents web search, page fetching, crawling, and caching with no API keys or cloud accounts.
  • Web Research Agents - Lets agents search many engines in parallel, read pages and gather cited, verifiable evidence for answering questions.
  • Agent Web Search and Page Reading - Gives coding agents tools to search the web, fetch pages and read their content through MCP, REST or embedded SDKs without API keys.
  • Autonomous Web Research Loops - Runs a time-budgeted loop that plans, searches, fetches, extracts, and synthesizes cited answers, logging each step and optionally shaping output to a schema.

AI search

Explore more awesome repositories

Describe what you need in plain English — the AI ranks thousands of curated open-source projects by relevance.

Start searching with AI
5,531 stars·459 forks·TypeScript·1 view
  • Autonomous Research Agents - Plans sub-queries, searches, fetches and synthesizes cited answers within a time budget, with a step log and optional schema-shaped output.
  • Verifiable Search Evidence - Returns each search result with a verbatim excerpt pinned to its source position, a quotable citation ID, and an inspectable score.
  • Web Page Markdown Converters - Loads a URL through a tiered router escalating from plain HTTP to a headless browser, returning clean markdown with metadata and links.
  • Offline Search Memory - Stores fetched and searched content locally so repeat queries return instantly and work offline without sending query data to the cloud.
  • Search Result Fusion Algorithms - Queries many engines in parallel, merges with rank fusion and reranking, and returns per-result scores with domain, time, phrase, and image filters.
  • Metasearch Engines - Queries many public engines in parallel, merges results with rank fusion and reranks on-device with an explainable relevance score.
  • Tiered Page Fetching - Escalates page fetching from plain HTTP to a headless browser on script-rendered, challenge, or thin pages and remembers what each site needs.
  • Agent Tool Server Connections - Exposes the same web tools through an agent-protocol server, REST endpoint, or embedded SDK so coding agents and apps can call them.
  • Agent Tool Servers - Serves the same web tools over MCP, REST with an OpenAPI contract, a command line and typed TypeScript and Python SDKs.
  • Local and On-Device Inference - Runs the browser engine and the ranking and embedding models on local hardware, removing per-query fees and cloud dependence.
  • Cited Brief Generators - Breaks a question into sub-queries, fetches many sources, and returns a cited report or structured brief for the host model to write from.
  • Source-Grounded Claim Verifiers - Returns exact quoted passages pinned to character-position spans in the source page so every claim can be verified against the original text.
  • Web Content Extractions - Turns pages into clean markdown or structured JSON with tables, metadata and named fields checked against the source text.
  • Search and Research - Queries public search engines through direct adapters with on-device reranking and embeddings, so searches cost nothing and need no cloud account.
  • Web Crawling and Scraping - Crawls public sites like a browser, honoring crawl rules by default and applying per-domain rate limits to keep requests polite.
  • Cached Content Search - Searches everything already retrieved with keyword or hybrid semantic matching and offers statistics, clearing, and change detection.
  • Web Scraping Result Caches - Stores retrieved page content in a local directory so repeated lookups avoid new network requests and stay private.
  • Web Page Caches - Stores fetched pages in a local keyword and vector index so repeat lookups and similarity searches skip refetching from the web.
  • Multi-Page Crawling - Follows links across a whole site using breadth-first, depth-first, sitemap, or map-only strategies, honoring crawl rules and rate limits.
  • Offline Caching - Stores every fetched and searched result in a local index on the user's machine so repeat lookups work offline and stay private.
  • Local Web Content Caches - Keeps fetched pages in a local keyword and vector index so repeat and similarity lookups run without refetching from the web.
  • Structured Data Extraction - Extracts tables, metadata, linked data markup, brand identity, and named or custom fields from pages as structured JSON.
  • Breadth-First Site Crawlers - Crawls sites breadth-first, depth-first, by sitemap, or map-only, honoring crawl rules, per-domain rate limits, and boilerplate removal.
  • Local-First Private Data Storage - Keeps cache, embeddings, models, and configuration on the user's machine and sends nothing to third parties unless an external language model is enabled.
  • Result Reliability Flags - Flags junk results with a dedicated scorer, reports failed engines, and labels stale cache so agents know how reliable data is.
  • Rank Fusion Algorithms - Queries many search engines at once, merges results with rank fusion, and reranks them with an on-device model into an explainable score.
  • Relevance Scoring - Breaks each result's relevance score into component factors for inspection and flags weak results as junk.
  • Blocked Page Detection - Labels stale cache, failed fetches, truncation, and bot-protected pages as failures rather than passing challenge pages off as real content.
  • Web Crawlers - Fetches pages as clean markdown, escalating from plain HTTP to a headless browser, while honoring crawl rules and per-domain rate limits.
  • Web Page Retrievers - Fetches pages through a tiered ladder from plain HTTP to a headless browser, learning per site and caching results locally.
  • Programmatic Page Interactions - Performs clicks, typing, scrolling, and screenshots on live pages and reuses signed-in sessions to reach content behind logins or dynamic interfaces.
  • Site Crawlers and Change Trackers - Crawls sites by breadth-first, depth-first or sitemap strategies with rate limits, and reports page changes to a webhook.
  • Web Data Ingestion - Local-first web intelligence and MCP server for extraction.

    • Star history

    Star history chart for knockoutez/wigoloStar history chart for knockoutez/wigolo

    How this analysis was created: This summary and feature list are AI-generated from collected project material and can contain mistakes. Stars, license and language are imported from GitHub. Inclusion does not mean that we have tested or audited this project. Check the source documentation for any feature you depend on. Learn more on our About page.

    Frequently asked questions

    What does knockoutez/wigolo do?

    Wigolo is a local-first web research toolkit for AI agents. It runs as a Model Context Protocol server that gives coding agents web search, page fetching, crawling, and caching without API keys or cloud accounts. Search queries many public engines in parallel, merges results with rank fusion, and reranks them with on-device models, attaching an explainable relevance score to each result.

    What are the main features of knockoutez/wigolo?

    The main features of knockoutez/wigolo are: Keyless Local Web Search, Web Research Agents, Agent Web Search and Page Reading, Autonomous Web Research Loops, Autonomous Research Agents, Verifiable Search Evidence, Web Page Markdown Converters, Offline Search Memory.

    Which projects share features with knockoutez/wigolo?

    Projects with overlapping indexed features include: any4ai/anycrawl — AnyCrawl is an AI-powered data extractor, automated web crawler, and headless browser orchestrator. It serves as a web… mendableai/firecrawl — Firecrawl is a headless browser automation tool and web crawling engine designed to extract structured data from the… deedy5/ddgs — ddgs is a metasearch engine and web content extractor that provides a toolkit for programmatically retrieving search… jina-ai/node-deepresearch — node-DeepResearch is an autonomous web research engine that uses large language models to iteratively search, read,… asciimoo/colly — Colly is a web scraping framework and concurrent crawler written in Go. It provides a system for traversing web pages,… eze-is/web-access — Web Access is a skill for local AI coding agents that gives them web access by driving a real Chromium-based browser.…

    Projects sharing features with Wigolo

    These projects share indexed features with Wigolo. Shared tags can include platform or build tooling; verify the primary use case before treating a result as a replacement.
    • any4ai/anycrawlany4ai avatar

      any4ai/AnyCrawl

      2,742View on GitHub↗

      AnyCrawl is an AI-powered data extractor, automated web crawler, and headless browser orchestrator. It serves as a web content extraction API and a gateway that connects crawling and scraping tools to language models using a standardized API protocol. The project specializes in converting unstructured website content into structured JSON or markdown optimized for AI assistants. It utilizes language models and JSON schemas to pull specific information into validated formats and provides capabilities for AI page summarization and LLM-optimized content extraction. The system manages comprehensi

      TypeScriptai-scrapingaitoolscrawl
      View on GitHub↗2,742
    • mendableai/firecrawlmendableai avatar

      mendableai/firecrawl

      139,399View on GitHub↗

      Firecrawl is a headless browser automation tool and web crawling engine designed to extract structured data from the web. It functions as an API that transforms raw website content and documents into clean markdown and JSON formats to serve as context for large language models. The project distinguishes itself by using natural language prompts to translate human instructions into targeted data extraction tasks and browser actions. It can execute interactive page navigation, such as clicking and scrolling, and perform automated web research to retrieve structured data without manual interventi

      TypeScript
      View on GitHub↗139,399
    • deedy5/ddgsdeedy5 avatar

      deedy5/ddgs

      2,754View on GitHub↗

      ddgs is a metasearch engine and web content extractor that provides a toolkit for programmatically retrieving search results from DuckDuckGo. It functions as a search API server and a Model Context Protocol server to integrate web search capabilities directly into large language model environments. The project distinguishes itself by aggregating text, image, news, and video results from multiple providers into a single interface. It includes a utility for fetching URLs and converting HTML content into markdown, plain text, or structured data. The system covers a broad range of search capabil

      Pythonapiddgsmcp
      View on GitHub↗2,754
    • jina-ai/node-deepresearchjina-ai avatar

      jina-ai/node-DeepResearch

      5,083View on GitHub↗

      node-DeepResearch is an autonomous web research engine that uses large language models to iteratively search, read, and reason over web content to answer complex questions. It provides a chat-based interface that displays real-time reasoning steps and final answers, and can be configured to focus exclusively on academic papers by limiting searches to academic repositories. The research engine operates through an agentic search-read-reason loop that repeatedly searches, reads, and reasons until a stopping condition is satisfied. It enforces a token budget to cap total consumption and failed at

      TypeScriptdeepresearchdeepsearch
      View on GitHub↗5,083
    Compare all 30 related projects→