Changelog

All notable changes to FetchtiumV2.

[2.0.1] - 2026-02-24 — Docs Sync for Current Snapshot

📝 Documentation Updates

Updated README and wiki pages to match current routes in src/app/api/**/route.ts.
Documented YouTube watch-url fast-path behavior in GET /api/v1/download and optional watch-url mode in GET /api/v1/merge.
Clarified that split-stream merge (videoUrl + audioUrl or videoH + audioH) remains supported and unchanged.

⚠️ Accuracy and Scope Notes

Removed/marked unavailable endpoint docs for routes not present in this snapshot (for example, /api/v1/events and /api/extract).
Added caveats that local Python wrapper mode and Docker/Railway full-profile claims depend on the presence of the Python api/ module, which is not included in this snapshot.
Kept deployment and platform coverage statements conditional on active profile and Python backend availability.

🔒 Tooling Notes

Documented dependency security remediation context (including minimatch override for eslint-config-next).
Updated lint/tooling notes to reflect current stack alignment (eslint@9 + eslint-config-next@16) without claiming warning-free lint output.

[2.0.0] - 2026-02-23 — YouTube Multi-Codec + UI Polish

🚀 Highlights

YouTube Multi-Codec Support - Returns ALL codecs per resolution (H.264, VP9, AV1) for user choice
Progressive Format Priority - 360p H.264 with audio now shown FIRST with "READY" status (no merge needed)
18 Video Sources - Returns 18 YouTube formats (vs 8 before) with proper codec grouping
DASH Warning Popups - Shows warning when downloading DASH formats requiring merge
New UI Banner - Added "Try our new version of social media downloader: DownAria" link below header

🎨 UI/UX Improvements

Logo Integration - Added icon.png logo in header using Next.js Image component
Auto-Extract on Paste - Paste button now automatically extracts when valid URL is in clipboard
Button Labels - Added text labels "paste" and "settings" to action buttons
Font Update - All text now uses JetBrains Mono font
Removed Backlinks - Cleaned up header area, removed backlinks section
Offline Message - Footer shows "When DownAria is offline, FetchtiumV2 is ready to wait" message
Section Rename - Changed "Supported Platforms" to "More media supported"

🔧 Technical Changes

Format Processing - process_youtube_formats() now separates progressive vs DASH, multi-codec per resolution
HLS Detection - Added /manifest/, index.m3u8, hls protocol checks
YouTube URL Handling - Added noplaylist=True and URL canonicalization for playlist/radio URLs
Codec Grouping - FormatList.tsx groups formats by codec: Progressive, H.264, VP9, AV1, Audio
Thumbnail Authorization - Fixed stream endpoint for YouTube thumbnails (i.ytimg.com, yt3.ggpht.com)
Download Speed - Added KB/s, MB/s indicator in loading state
Merge Optimization - Merge endpoint now uses audio-copy flag and reconnect flags

🐛 Fixes

Python Bytecode Cache - Documented cache clearing with taskkill /F /IM python.exe to avoid stale results
CookieModal Animation - Fixed close behavior to animation-driven close lifecycle
DownAria Link Update - Changed from downaria.site to downaria.vercel.app

📦 Files Changed

~ package.json                    # Version 2.0.0
~ src/app/layout.tsx             # JetBrains Mono font
~ src/app/page.tsx               # Logo, banner, button labels, section rename
~ src/components/ExtractForm.tsx # Auto-extract on paste, text labels
~ src/components/FormatList.tsx  # Codec grouping, warning popups
~ api/services/formats.py        # Multi-codec, progressive priority
~ api/routes/extract.py           # URL canonicalization

[1.8.0] - 2026-02-23 — FetchtiumV2 Modular Restructure

🚀 Highlights

Introduced modular Python API architecture with app factory and blueprints.
Added extractor service layer modules (formats, resolver, ytdlp, gallery_dl, transforms).
Added TypeScript extractor architecture split (native, wrappers, shared types, Python client).
Added new Next.js route POST /api/extract while preserving /api/v1/extract compatibility.
Standardized Python runtime port to 5000 across scripts, env, and orchestration.
Fixed CookieModal close behavior to animation-driven close lifecycle.

🧪 Testing

Added Python tests for config, security, errors, formats, routes, wrappers, resolver, and smoke checks.
Added TypeScript unit tests with Vitest for Python client, wrappers, router detection, and health route.

📦 Ops and Docs

Added docker-compose.yml for local dual-service setup.
Simplified start.sh dual-process orchestration (Python background + Next foreground).
Updated docs and added architecture/migration documentation.

[1.7.0] - 2026-02-23 — Dynamic Deployment Profile

🚀 What's New

Dynamic Deployment Profile - Single codebase works on both Vercel and Docker/Linux
- vercel profile: Native extractors only (Facebook, Instagram, TikTok, Twitter, Pixiv)
- full profile: Native + Python extractors (YouTube, BiliBili, SoundCloud, etc.)
Auto Environment Detection - Automatically detects Vercel vs Docker/Railway
- Priority: EXTRACTOR_PROFILE env var → Vercel auto-detect → Default to full
Platform Availability Error - New PLATFORM_UNAVAILABLE_ON_DEPLOYMENT error code
- Returns helpful message when Python platforms requested on Vercel deployment
Public Access Model - Removed API key requirement, open access for all users

🔧 Technical Changes

EXTRACTOR_PROFILE env var - Controls extractor capability at runtime
getExtractorProfile() - Resolves profile from environment with fallback chain
isPythonEnabled() - Helper to check Python extractor availability
Direct HTTP calls - Removed Next.js rewrites, Python service called directly
PYTHON_API_URL support - Configurable Python endpoint for non-Vercel deployments

📦 Files Changed

+ .gitattributes                    # LF enforcement for shell scripts
~ vercel.json                       # Added framework, regions, EXTRACTOR_PROFILE
~ next.config.ts                    # Removed rewrites, updated CORS headers
~ src/lib/config/index.ts           # Added getExtractorProfile(), isPythonEnabled()
~ src/types/config.ts               # Added extractorProfile, removed apiKeys
~ src/app/api/v1/extract/route.ts   # Profile-aware routing, platform blocking
~ src/lib/utils/error.utils.ts      # Added PLATFORM_UNAVAILABLE_ON_DEPLOYMENT
~ Dockerfile                        # Added EXTRACTOR_PROFILE=full, use requirements.txt
~ start.sh                          # Added EXTRACTOR_PROFILE env var
~ src/app/page.tsx                  # Logo from public/icon.png, Lucide icons
~ src/app/changelog/page.tsx        # Logo from public/icon.png, Lucide icons
- StatusBadge component             # Removed cold/warm state from UI

[1.6.0] - 2026-01-13 — Go Backend Extractors Reorganization

🔧 Refactoring (Go Backend)

Extractors Reorganization - Restructured internal/extractors/ into logical subfolders:
- core/ - Shared types and interfaces (types.go)
- registry/ - Platform detection and extractor registry (registry.go, patterns.go)
- cookies/ - Cookie parsing and management (cookies.go)
- native/ - TypeScript-equivalent extractors (Facebook, Instagram, TikTok, Twitter, Pixiv)
- aria-extended/ - yt-dlp/gallery-dl wrapper extractors (YouTube, BiliBili, SoundCloud, etc.)
- tests/ - Property-based and integration tests
HTTP Package Simplification - Merged and cleaned up pkg/http/:
- client.go - HTTP client with connection pooling
- pool.go - Connection pool management
- helpers.go - Request/response utilities
Type Fixes - Fixed ExtractorFactory type mismatch between registry and extractors

📦 Files Changed (Go)

.planning/fetchtium_go/internal/extractors/
├── core/types.go           # Shared types (ExtractResult, Source, etc.)
├── registry/
│   ├── registry.go         # Extractor registration & lookup
│   └── patterns.go         # URL pattern matching
├── cookies/cookies.go      # Cookie parsing (Netscape, JSON, raw)
├── native/                 # Native extractors
│   ├── facebook/
│   ├── instagram/
│   ├── tiktok/
│   ├── twitter/
│   └── pixiv/
├── aria-extended/          # yt-dlp/gallery-dl wrappers
│   ├── wrapper.go
│   ├── youtube/
│   ├── bilibili/
│   └── ...
└── tests/                  # Test files

.planning/fetchtium_go/pkg/http/
├── client.go               # Simplified HTTP client
├── pool.go                 # Connection pool
└── helpers.go              # Utilities

[1.5.1] - 2026-01-12 — Pixiv Native Support + Hotfixes

🚀 What's New

Pixiv Native Extractor - Direct API integration without gallery-dl dependency
- Single and multi-page artwork support
- Original quality images with file sizes
- Automatic Referer header handling via proxy
All Resolutions Support - No more 360p/480p/720p/1080p limit
- Returns ALL available video qualities (144p to 4K+)
- Deduplication by exact height (one format per resolution)
Changelog Page - New /changelog page that reads directly from CHANGELOG.md
- Auto-updates when changelog is modified
- Styled markdown rendering
Format Badge - Shows original format with codec in FormatList
- Combined format: [HLS-H.264], [MP4-VP9], [DASH-AV1], [Opus]
- Mime type shown below for output format

🔧 Technical Changes

src/lib/extractors/pixiv/ - New native TypeScript extractor using Pixiv's /ajax/illust/{id} API
Removed HTTP/2 from TypeScript - Simplified to undici only (Python keeps httpx HTTP/2)
Proxy support for Pixiv - Stream/download/thumbnail routes add Referer: https://www.pixiv.net/
needsProxy flag - Pixiv sources marked for automatic proxy routing

🐛 Fixes

PlayerModal audio leak - Fixed audio continuing after modal close (proper cleanup with load())
Merge loading message - Changed to "Merging video and audio for playback... please wait"
FormatList Open button - Now uses proxy for needsProxy sources (Pixiv images)
ResultCard thumbnails - Added Pixiv to proxy thumbnail platforms

📦 Files Changed

+ src/lib/extractors/pixiv/index.ts    # New Pixiv native extractor
+ src/app/changelog/page.tsx           # Changelog page
+ src/app/api/changelog/route.ts       # Changelog API endpoint
- src/lib/core/network/http2-client.ts # Removed (Python has own httpx)
~ src/lib/core/network/client.ts       # Simplified to undici only
~ src/lib/extractors/index.ts          # Register PixivExtractor
~ src/lib/extractors/python-platforms.ts # Remove Pixiv from Python
~ src/app/api/v1/stream/route.ts       # Add Pixiv Referer
~ src/app/api/v1/download/route.ts     # Add Pixiv Referer
~ src/app/api/v1/thumbnail/route.ts    # Add pximg.net + Referer
~ src/components/PlayerModal.tsx       # Fix audio cleanup
~ src/components/FormatList.tsx        # Format badge + proxy for needsProxy
~ src/components/ResultCard.tsx        # Pixiv thumbnail proxy
~ src/app/page.tsx                     # Add changelog link in header
~ src/middleware.ts                    # Add /api/changelog to public routes
~ api/extract.py                       # Remove Pixiv, all resolutions support

[1.5.0] - 2026-01-11 — HTTP/2 Support + Rule34Video + Stream Improvements

🚀 What's New

HTTP/2 Support (TypeScript) - Auto-detect and use HTTP/2 for HTTPS requests when available (Vercel Functions, Railway)
HTTP/2 Support (Python) - Shared httpx client with HTTP/2 and 5-minute connection pooling
Rule34Video Support - Multi-format extraction (360p-4K) via yt-dlp with URL resolution
Eporner Multi-Format - Returns all available formats (240p-1080p 60fps) with file sizes
URL Resolution - Wrapper URLs (Rule34Video, Eporner) resolved to real CDN URLs in parallel
Stream Expired URL Detection - Returns "URL_EXPIRED" error instead of following rickroll redirects

🔧 Technical Changes

http2-client.ts - New HTTP/2 client with session pooling (5-min keepalive), auto-cleanup
client.ts - Smart detection: HTTP/2 for HTTPS when available, fallback to undici HTTP/1.1
api/extract.py - Shared HTTP_CLIENT with httpx HTTP/2, 5-min keepalive, 10 max connections
resolve_media_url() - Resolves wrapper URLs to CDN, removes download=true params
resolve_media_urls_parallel() - Parallel URL resolution with ThreadPoolExecutor (5 workers)
transform_eporner_result() - Checks quality field first (Rule34Video uses this), adds "p" suffix
stream route - Detects rickroll/placeholder redirects and returns 410 Gone with URL_EXPIRED code

🐛 Fixes

Rule34Video Quality - Fixed quality showing "0", "1", "2" instead of "360p", "720p", "1080p"
Rule34Video Playback - Fixed "Open in new tab" redirecting to rickroll (URL resolution)
Rule34Video Download - Fixed browser downloading instead of playing (removed download=true param)
Eporner Quality - Fixed quality string formatting for all formats

📦 New Files

src/lib/core/network/http2-client.ts  # HTTP/2 client with session pooling

📋 Dependencies

# Python (requirements.txt)
httpx[http2]  # HTTP/2 support for Python extractor

🔧 Environment Detection

// Auto-detected at startup (zero runtime overhead)
- Vercel Functions → HTTP/2 ✓
- Railway Node.js → HTTP/2 ✓
- Vercel Edge → undici HTTP/1.1 (no http2 module)
- Local dev → HTTP/2 ✓

[1.4.0] - 2026-01-11 — Twitter, Instagram Stories, Pinterest Video, YouTube HLS

🚀 What's New

Twitter/X Cookie Support - Fixed Netscape cookie format parsing with netscapeToCookieHeader() converter
Twitter TweetWithVisibilityResults - Handle GraphQL wrapper for tweets with visibility restrictions
Instagram Stories - Full support via Internal API, fetches all story items as carousel
Instagram Reels URL - Support /reels/ (plural) in addition to /reel/
Pinterest Video - Extract MP4 sources from video pins (V_720P, V_HLSV4 fallback)
YouTube HLS Proxy - True streaming playback via HLS.js instead of full download
YouTube HLS Download - FFmpeg routes through proxy for proper segment fetching

🔧 Technical Changes

New Endpoint - /api/v1/hls-proxy proxies HLS manifest and segments on-demand
Thumbnail Endpoint - Changed from hash-based (?h=) to direct URL (?url=) for easier API usage
CookieModal - Added netscapeToCookieHeader() to convert Netscape format to HTTP Cookie header
Twitter extract.ts - Unwrap TweetWithVisibilityResults to get actual tweet from result.tweet
Instagram scanner.ts - Added fetchStory(), fetchStoryByUsername() functions
Instagram extract.ts - Added parseStoryResponse(), handle reels[userId] response structure
Instagram patterns - Added XDTGraphSidecar, XDTGraphImage, XDTGraphVideo typename checks
api/extract.py - transform_pinterest_result() handles video pins with contentType: 'video'
PlayerModal - Split needsHlsStream into needsHlsTranscode (Opus) and needsHlsProxy (YouTube)
hls-stream route - Detects YouTube HLS and routes through proxy for FFmpeg download
middleware.ts - Added /api/v1/hls-proxy to public routes and streaming endpoint checks

🐛 Fixes

Twitter Cookies - Fixed cookie parsing not working with Netscape export format
Twitter Visibility - Fixed extraction failing on tweets wrapped in TweetWithVisibilityResults
Instagram Stories - Fixed "reels_media[0] undefined" error - API returns reels[userId] not reels_media
Instagram Carousel - Fixed typename check missing XDT-prefixed types from Internal API
YouTube HLS Play - Fixed video downloading entire stream before playing (now true streaming)
YouTube HLS Download - Fixed 403 errors by adding Sec-Fetch-* headers to proxy requests

📦 New Files

src/app/api/v1/hls-proxy/route.ts  # HLS proxy for CORS-restricted streams

[1.3.0] - 2026-01-11 — Extract.py Cleanup + API Docs + UI Polish

🚀 What's New

YouTube Loading Indicator - Shows "Processing stream... please wait" during FFmpeg transcoding
BiliBili Loading Indicator - Shows "Merging video and audio... please wait" during merge
Thumbnail Proxy - BiliBili and YouTube thumbnails now display correctly via stream proxy

🎨 UI Improvements

Header - Added "API Docs" link button
Footer Redesign - Clean 2-column grid layout with Links (left) and Platforms (right)
Platform Badges - Color-coded: Native (green), Python (blue), NSFW (red)
Deploy Links - Railway marked "Full Support", Vercel marked "Limited"

🔧 Extract.py Cleanup (Major Refactor)

Centralized Configuration - Single DEFAULT_USER_AGENT and DEFAULT_HEADERS constants
Platform Registry - New PLATFORM_CONFIG dict with extractor type, patterns, NSFW flag, custom headers
Derived Lists - YTDLP_PLATFORMS, GALLERY_DL_PLATFORMS, NSFW_PLATFORMS auto-generated from config
Short URL Resolution - Unified resolve_short_url() for pin.it, b23.tv, redd.it
Format Processing - New process_video_formats() and process_audio_formats() functions
Header Merge - merge_headers() helper for platform-specific header overrides
File Organization - Clear sections: Imports, Constants, Config, Security, Detection, Extractors, Processing, Transformers, Routes

🐛 Fixes

Thumbnail Display - Fixed BiliBili/YouTube thumbnails not showing (removed hash-based approach, use stream proxy)
YouTube Playback - Fixed immediate loading without indicator
Middleware Cleanup - Removed all console.log debug statements

🔧 Technical Changes

ResultCard - Uses /api/v1/stream?url=... for thumbnail proxy instead of hash
PlayerModal - Added needsHlsStream to useEffect dependencies
extract route - Simplified URL storage, removed thumbnailHash assignment

[1.2.4] - 2026-01-11 — BiliBili Support + JSON Cleanup

🚀 What's New

BiliBili Video+Audio Merge - Downloads now include audio! FFmpeg merges video and audio streams from BiliBili DASH segments
BiliBili Playback - Video playback with audio via server-side FFmpeg merge
Cleaner JSON Response - Removed hash from source objects to reduce payload size

🐛 Fixes

BiliBili Download - Fixed video downloads being video-only (no audio)
BiliBili Playback - Fixed playback having no audio by merging video+audio in FFmpeg
BiliBili Thumbnail - Added bstarstatic.com to stream endpoint Referer check

🔧 Technical Changes

hls-stream endpoint - Added audioUrl parameter for BiliBili video+audio merge (2 FFmpeg inputs)
FormatList - Passes audio URL for BiliBili video downloads
PlayerModal - Includes audioUrl in hls-stream request for BiliBili
extract route - Removed hash from source objects (URLs still stored for validation)

[1.2.3] - 2026-01-11 — YouTube HLS Video Playback Fix

🐛 Fixes

YouTube Video Playback - Fixed video playing without audio. FFmpeg now includes audio track (-c:a aac) instead of stripping it (-an)
Separate Audio Logic - Fixed FormatList incorrectly passing separate audio for YouTube HLS streams. Now only uses separate audio when hasAudio === false explicitly
Video Sizing - Video player now fits modal properly with max-h-[80vh] object-contain

🔧 Technical Changes

hls-stream endpoint - Video output now remuxes both video and audio from HLS source
FormatList - Simplified audio detection logic, removed URL-based guessing
PlayerModal - Cleaned up console.log statements

[1.2.2] - 2026-01-10 — HLS Opus Streaming + YouTube

🚀 What's New

YouTube Support - Video extraction via yt-dlp with quality selection, codec info, fps
HLS Opus Playback - SoundCloud Opus HLS streams now play in browser via server-side FFmpeg transcoding
Experimental Badge - HLS and Opus formats marked with ⚡ Experimental badge
Audio Preview - Thumbnail preview with "🎵 Audio Only" badge for audio playback
New Endpoint - /api/v1/hls-stream converts HLS to progressive MP3 stream

🔧 Technical Changes

FFmpeg Integration - Uses ffmpeg-static package for server-side transcoding
Path Resolution Fix - Fixed FFmpeg binary path in Next.js bundled environment
Middleware Update - Added /api/v1/hls-stream to public routes
YouTube Metadata - Includes codec (H.264/VP9/AV1), fps, hasAudio flags
YouTube Simplified Output - 1 format per resolution (max 4: 1080p, 720p, 480p, 360p), priority: hasAudio > H.264 > VP9 > AV1
YouTube Audio Simplified - Only 2 audio formats: AAC ~128kbps + Opus ~128kbps

📋 New Dependencies

{
  "ffmpeg-static": "^5.3.0"
}

[1.2.1] - 2026-01-10 — Python Extractors

🚀 What's New

Python Extractors - Added support for 6 new platforms via yt-dlp and gallery-dl
SoundCloud - Audio extraction with multiple formats (MP3, AAC, Opus)
BiliBili - Video extraction with multiple qualities (144p-720p)
Reddit - Video/image/gallery extraction
Pixiv - Artwork extraction (NSFW support with cookies)
Eporner - Adult video extraction (NSFW)
Rule34Video - Adult video extraction (NSFW)

🔧 Technical Changes

Vercel Python Functions - Flask-based serverless functions alongside Next.js
Hybrid Architecture - TypeScript extractors (Twitter, Instagram, TikTok, Facebook) + Python extractors (new platforms)
Unified API - All platforms accessible via same /api/v1/extract endpoint
NSFW Flag - Response includes isNsfw: true for adult platforms
Concurrent Dev - npm run dev runs both Next.js and Python Flask server

📦 New Files

api/py/
└── extract.py         # Flask-based Python extractor

src/lib/extractors/
└── python-platforms.ts  # Platform detection for Python routing

📋 New Dependencies

# Python (requirements.txt)
yt-dlp
gallery-dl
flask

# Node (devDependencies)
concurrently

🔧 Config Changes

vercel.json - Python 3.12 runtime configuration
next.config.ts - Dev rewrites for Python server proxy
package.json - New dev scripts with concurrently

[1.2.0] - 2026-01-10 — Next.js Migration + Streaming Fixes

🚀 What's New

Next.js 16 with App Router and Turbopack
TypeScript full codebase conversion with strict types
React frontend with custom hooks (useExtract, useStatus)
Server Components for optimal performance
5-minute warm status - extended keep-alive timeout from 30s to 5min
HLS.js Integration - Better audio/video playback support

🔧 Streaming & Download Fixes

Stream Mode - Added streamMode option to disable body timeout for large files
Progressive Streaming - Audio/video now streams progressively instead of buffering entire file
HLS Support - Added HLS.js library for m3u8 playlist playback
Better Error Handling - PlayerModal shows specific error codes (MEDIA_ERR_DECODE, etc.)
MIME Mappings - Comprehensive audio MIME type mappings (opus, ogg, flac, wav, aac)

🎵 Python Extractor Improvements

Extension Field - Source objects now include extension field
Filename Generation - Auto-generated filenames: Author_Title_Quality.ext
Audio Metadata - Bitrate and filesize fields for audio sources
Sanitized Output - Proper filename sanitization for special characters

🔄 What's Different

Framework: Fastify → Next.js 16 App Router
Language: JavaScript → TypeScript
Frontend: Vanilla JS → React components
Logging: Pino → Console-based logger (simplified)
SSE: Custom implementation → ReadableStream API
Middleware: Fastify hooks → Next.js middleware.ts
Cookie handling: Frontend no longer auto-sends cookies, backend decides

🐛 Fixes

File size detection - Fixed undici maxRedirections not supported error
Image extraction - Removed overly aggressive skip patterns (.webp, _s\d+x\d+) that blocked valid Facebook images
Content type detection - Fixed permalink.php URLs not detected as posts
Filename format - Updated to Author_Description_Quality_[DownAria].ext
Playback timeout - Fixed audio/video timeout on large files (bodyTimeout: 0 for streams)

📦 New Components

src/components/
├── ExtractForm.tsx    # URL input with paste button
├── ResultCard.tsx     # Metadata display
├── FormatList.tsx     # Download options
├── PlayerModal.tsx    # Video/audio player with HLS.js
├── CookieModal.tsx    # Cookie management
├── StatusBadge.tsx    # SSE warm/cold indicator
└── JsonOutput.tsx     # Collapsible JSON with copy

📋 New Dependencies

{
  "hls.js": "^1.6.15"  // HLS streaming support
}

🔧 API Changes

No breaking changes - all endpoints remain the same:

POST /api/v1/extract
GET /api/v1/stream
GET /api/v1/download
GET /api/v1/status
GET /api/v1/events
GET /api/health

[1.0.0] - 2026-01-8 — Pre-Migration Release (Fastify)

🎉 Initial Complete Release

This is the complete Fastify-based version before migrating to Next.js.

Core Architecture

Fastify web framework with streaming support
Undici HTTP client with connection pooling
Pino structured logging system
Environment-based configuration (.env)
ES Modules (type: module)

Facebook Extractor

Full support: Videos, Reels, Stories, Posts, Galleries, Groups
Cookie authentication (Netscape + JSON + raw string formats)
Smart retry strategy: guest first → cookies on auth error
Stories always use cookies (required)
Engagement stats extraction from JSON + title fallback
Content issue detection (age-restricted, private, deleted, expired)
URL resolution for short links (fb.watch, fb.me)

API Endpoints (v1)

POST /v1/extract - Main extraction with cookie support
GET /v1/stream - Video/audio streaming proxy (range requests)
GET /v1/download - File download proxy
GET /v1/status - Server status
GET /v1/events - SSE real-time status updates
GET /health - Health check

Security

Path traversal protection - canonical path validation, encoding bypass prevention
SSRF protection - internal IPs, metadata endpoints, numeric/octal/hex IP blocking
XSS sanitization - using xss library
SQL injection patterns - defense in depth
CRLF/header injection - blocked
Request size limit - 1MB max
Rate limiting - 100 req/min per IP
Security headers - via @fastify/helmet
Frontend sanitization - backticks, template literals blocked

Cookie System

Priority: Client cookie → Server cookie (future) → File cookie
Auto-detect format: JSON array, JSON object, Netscape, raw string
parseCookies() universal parser
getFacebookCookies({ clientCookie }) with priority handling

Response Builder

Standardized API response format
meta object: responseTime, accessMode, publicContent
stats normalization for all platforms
buildResponse(), buildErrorResponse() helpers

Filename Generation

Universal addFilenames() at route level
Format: author_contentType_title(25)_quality.ext
Unicode support (Chinese, Japanese, etc.)
Uses getExtensionFromMime() helper

Self-Registering Extractors

Each extractor defines static patterns and static match()
Registry auto-discovers extractors
url.utils.js is platform-agnostic

Frontend (Vanilla JS)

Dark theme with Tailwind CSS (CDN)
Extract form with paste button
Result card with thumbnail, metadata, stats
Format list with Play, Open, Download buttons
Player modal (video/audio) using proxy stream
Cookie modal with platform tabs (Facebook, Instagram, TikTok)
Copy JSON button
SSE status indicator (Warm/Cold)
Download All with progress
Input sanitization (XSS, command injection)

Output Format

{
  "success": true,
  "platform": "facebook",
  "contentType": "reel",
  "title": "...",
  "author": "...",
  "stats": { "views": 9700, "likes": 340 },
  "items": [{
    "index": 0,
    "type": "video",
    "thumbnail": "...",
    "sources": [{
      "quality": "hd",
      "url": "...",
      "resolution": "1280x720",
      "mime": "video/mp4",
      "size": 4493909,
      "filename": "author_reel_title_hd.mp4"
    }]
  }],
  "meta": {
    "responseTime": 2049,
    "accessMode": "public",
    "publicContent": true
  }
}

Dependencies

fastify, @fastify/cors, @fastify/helmet, @fastify/static
undici, node-html-parser, xss, validator
pino, pino-pretty, dotenv

Placeholders (Not Implemented)

YouTube extractor
Instagram extractor
TikTok extractor
Twitter/X extractor

Pre-1.0.0 Development Notes

2026-01-10 — Session Summary

Removed legacy /api/* routes (replaced by /v1/*)
Created /v1/stream with range request support
Added player modal for video/audio playback
Implemented SSE for status updates (replaced polling)
Self-registering extractors pattern
Comprehensive security hardening
Cookie system with client upload support
Response builder with meta/stats
Filename generation utility
Stats extraction with title fallback
Copy JSON button fix (CSP compliance)

FilesExpand file tree

CHANGELOG.md

Latest commit

History

CHANGELOG.md

File metadata and controls

Changelog

[2.0.1] - 2026-02-24 — Docs Sync for Current Snapshot

📝 Documentation Updates

⚠️ Accuracy and Scope Notes

🔒 Tooling Notes

[2.0.0] - 2026-02-23 — YouTube Multi-Codec + UI Polish

🚀 Highlights

🎨 UI/UX Improvements

🔧 Technical Changes

🐛 Fixes

📦 Files Changed

[1.8.0] - 2026-02-23 — FetchtiumV2 Modular Restructure

🚀 Highlights

🧪 Testing

📦 Ops and Docs

[1.7.0] - 2026-02-23 — Dynamic Deployment Profile

🚀 What's New

🔧 Technical Changes

📦 Files Changed

[1.6.0] - 2026-01-13 — Go Backend Extractors Reorganization

🔧 Refactoring (Go Backend)

📦 Files Changed (Go)

[1.5.1] - 2026-01-12 — Pixiv Native Support + Hotfixes

🚀 What's New

🔧 Technical Changes

🐛 Fixes

📦 Files Changed

[1.5.0] - 2026-01-11 — HTTP/2 Support + Rule34Video + Stream Improvements

🚀 What's New

🔧 Technical Changes

🐛 Fixes

📦 New Files

📋 Dependencies

🔧 Environment Detection

[1.4.0] - 2026-01-11 — Twitter, Instagram Stories, Pinterest Video, YouTube HLS

🚀 What's New

🔧 Technical Changes

🐛 Fixes

📦 New Files

[1.3.0] - 2026-01-11 — Extract.py Cleanup + API Docs + UI Polish

🚀 What's New

🎨 UI Improvements

🔧 Extract.py Cleanup (Major Refactor)

🐛 Fixes

🔧 Technical Changes

[1.2.4] - 2026-01-11 — BiliBili Support + JSON Cleanup

🚀 What's New

🐛 Fixes

🔧 Technical Changes

[1.2.3] - 2026-01-11 — YouTube HLS Video Playback Fix

🐛 Fixes

🔧 Technical Changes

[1.2.2] - 2026-01-10 — HLS Opus Streaming + YouTube

🚀 What's New

🔧 Technical Changes

📋 New Dependencies

[1.2.1] - 2026-01-10 — Python Extractors

🚀 What's New

🔧 Technical Changes

📦 New Files

📋 New Dependencies

🔧 Config Changes

[1.2.0] - 2026-01-10 — Next.js Migration + Streaming Fixes

🚀 What's New

🔧 Streaming & Download Fixes

🎵 Python Extractor Improvements

🔄 What's Different

🐛 Fixes

📦 New Components

📋 New Dependencies

🔧 API Changes

[1.0.0] - 2026-01-8 — Pre-Migration Release (Fastify)

🎉 Initial Complete Release

Core Architecture