$ show miniflux/v2
Miniflux is a self-hosted, minimalist RSS/Atom feed reader implemented as a monolithic Go web application. It relies exclusively on PostgreSQL for persistence and serves a server-rendered web UI, a comprehensive REST API, and compatibility endpoints for the Fever and Google Reader APIs. The application is designed for low resource consumption and easy deployment, evidenced by its single-binary nature and extensive packaging (Docker, RPM, Deb, systemd). The core strength of the codebase lies in its feed ingestion pipeline, which demonstrates exceptional defensive programming against the chaotic and non-standardized world of web feeds. This pipeline includes robust format detection (RSS, Atom, RDF, JSON Feed), character encoding detection for legacy charsets, a port of the Mozilla Arc90 readability algorithm for content extraction, a highly restrictive HTML sanitizer with SSRF protection, and a DSL-based content rewrite engine. The system supports multi-user authentication (local, OAuth2, WebAuthn), extensive third-party integrations (Discord, Slack, Telegram, Pinboard, etc.), and background job processing for feed refreshes. The codebase exhibits high craftsmanship with clean abstractions, consistent builder patterns, and thorough testing (unit, integration, fuzz). Several components, such as the SSRF-safe HTTP client builder, the readability extraction algorithm, the i18n pluralization engine, and the HTML sanitizer, are isolated, well-documented, and could be extracted for use in other Go projects. The project is actively maintained with modern Go features, zero Dependabot alerts, and a strong CI/CD pipeline. The main maintenance burden is the sprawl of third-party integrations and the monolithic JavaScript frontend, but the core Go server code is exceptionally well-engineered.
The LLM identified 10 files worth highlighting for reuse:
A standalone port of the Mozilla Arc90 readability algorithm for extracting primary article content from HTML. It uses candidate scoring, link density weighting, and sentence detection. It is well-tested, isolated, and could be reused in any Go project needing article extraction.
A fluent HTTP response builder supporting brotli/gzip/deflate compression, ETag/If-None-Match caching, security headers (CSP, X-Frame-Options), and Content-Disposition escaping. It is a self-contained utility demonstrating a clean builder pattern for HTTP responses.
A thorough HTML sanitizer using tag/attribute allowlists, iframe domain allowlists (YouTube, Vimeo, etc.), tracking parameter stripping, 1x1 pixel tracker detection, and depth limits. It is a critical security component that is well-isolated and handles untrusted HTML from the web robustly.
A fluent HTTP client builder with built-in SSRF protection via private-network blocking at the dialer level. Supports proxies, TLS configuration, redirect control, and custom headers. The SSRF protection pattern is sophisticated and reusable.
An i18n system with lazy-loaded JSON translations, a sophisticated pluralization engine supporting complex rules (Russian, Arabic, Polish), and a LocalizedError pattern for user-facing errors. The pluralization logic is a reusable gem.
A DSL-based content rewrite engine that parses and executes rules like replace("a"|"b"), add_youtube_video, and fix_medium_images. The rule parsing and execution engine is self-contained and demonstrates a clean DSL pattern.
A robust date parser designed for the chaotic world of RSS/Atom feeds. Handles a vast array of date formats, timezone abbreviations, and localized month names. Includes fuzz testing and is a prime candidate for extraction into a standalone library.
A character encoding detection and decoding module for HTML/XML. Handles legacy charsets (ISO-8859-1, KOI8-R, Windows-1251) and edge cases like meta tags appearing after 1024 bytes. Solves a common data cleaning problem robustly.
An entry filtering rules engine supporting blocklist/keeplist patterns with regex and date-based predicates (before:, after:, max-age:). The rule parsing and evaluation logic is clean and reusable.
A generic worker pool for executing background jobs. Manages concurrency, a job queue, and error handling. It is well-tested and demonstrates a clean implementation of a standard concurrency pattern.