Marreta

Marreta is an open-source, self-hosted web content proxy with over 1,100 GitHub stars and active development by the Manual do Usuário team in Brazil. It fetches web pages and processes them to produce clean, readable versions stripped of tracking cookies, advertising, cookie consent popups, and HTML noise. The name means "hammer" in Portuguese — it is designed to break through the barriers that prevent direct access to content. Built on PHP 8.4 with Nginx, it runs as a single lightweight container backed by an embedded SQLite cache, with no external database dependency.

Common Use Cases

Researchers and journalists use Marreta to access articles from sources with metered or cookie-based access gates, getting clean text content without being tracked, profiled, or required to create an account. Privacy-focused readers route news links through Marreta to strip tracking parameters and read articles without sharing their identity with the publisher's analytics stack. Developers integrate the JSON API into RSS aggregators, read-it-later pipelines, and browser automation workflows to extract clean article content programmatically. Self-hosters deploy Marreta as a household reading proxy, giving family members a shared instance that caches frequently read sources and avoids re-fetching the same articles repeatedly. Teams maintaining internal knowledge bases use it to archive snapshots of external articles at a point in time, preserving content that might later be moved behind a harder paywall or deleted.

Key Features

• URL cleaning and normalization: strips tracking parameters, UTM tags, and redirect wrappers before fetching
• Cookie-free page retrieval with customizable user agent and header manipulation per domain
• HTML optimization: removes ads, scripts, popups, banners, and non-content elements from page output
• Local SQLite-backed caching with configurable retention and automatic scheduled cleanup
• Per-domain custom rules for CSS removal, JavaScript injection, and specific cookie or header overrides
• Dual access: rendered HTML page at /p/{url} and structured JSON response at /api/{url}
• Optional Selenium Hub integration for JavaScript-heavy pages that require a real browser renderer
• Optional S3-compatible storage backend for distributed or high-volume cache deployments
• Optional proxy list support for routing requests through external proxies
• Multi-language interface: English, Portuguese, Spanish, German, and Russian

Why deploy Marreta on Hostinger VPS

Cloud-based reader proxies and paywall bypass services process your reading requests on their own servers — meaning a third party sees every article you read, every link you follow, and every domain you access. Running Marreta on a Hostinger VPS keeps all of that browsing activity on infrastructure you control, with no usage logs sent to an external service. A VPS provides the persistent availability and outbound bandwidth that a caching proxy requires: unlike a home server, it is always online and reachable without port-forwarding or dynamic DNS. The built-in cache dramatically reduces repeated bandwidth consumption — once an article is fetched and cleaned, subsequent reads are served instantly from local storage. With PHP-FPM OPcache optimizations baked into the container and Traefik handling HTTPS termination, the deployment runs efficiently even on a small VPS tier, making it a practical and cost-effective private alternative to hosted article-cleaning services.