EFVEFV
Sign in
EFV CHANGELOG

Change log

Document every improvement and stay up to date on EFV’s latest advancements.

  1. Added EFV parsing video player, which prevents mobile browser hijacking, supports remembering playback, progress bar hover preview, splash ads, pause ads, pop-up ads, text watermark, image watermark and scrolling text ads, and is adapted to HLS double encryption hotlink protection. That is, once this hotlink protection is enabled, you must go through this video player or implement the decryption logic yourself on the front end in order to playback the m3u8 video of EFV. Details: Click to view.
  2. Added automatic short-drama episode splitting, which automatically detects the episode boundaries of merged long videos, with visual fine-tuning on the review page (drag to extract frames, union adjustment); after confirmation, it splits within seconds database ingestion and automatically enters the transcoding queue. It can also automatically bind episodes, i.e., once transcoding is completed, they are automatically added to the series and sorted by episode number.
  3. Optimized the recognition of M3U8 interstitial ads in the download center, and added interstitial ad retention and playback: the ad segmentation identified during ad-free download is automatically merged into mp4 for retention, and the download center allows one-click playback for checking, with automatic cleanup upon expiration.
  4. Added 9 movement trajectories to the image moving watermark (such as 8-shaped orbiting, Z-shaped snake, irregular drifting, etc.), which you can choose freely.
  5. Optimized FTP multi-user upload, covers and detail json no longer need to be uploaded before the video, i.e., covers and detail json are automatically filled in or updated whether they are uploaded first or later.
  6. Optimized the burning of intro videos and cover ads, allowing separate intro videos and cover ads to be set for landscape and portrait videos respectively.
  7. Optimized the cloud storage sync feature, supporting deletion of only the specified category data after sync is complete.
  8. Fixed the issue where some fonts did not take effect when selecting built-in fonts for ASS/SRT subtitles burning, and added a custom font option so you can use your own font style.
  9. Fixed the issue where the automatic subtitles extraction feature did not take effect for a few srt formats.
  10. Some feature optimizations and BUG fixes.
  1. Optimized the route distribution management feature and added route settings for international and Hong Kong, Macao, and Taiwan; non-mainland-China IP users will use this route by default when watching videos.
  2. Optimized the IP access restriction setting, allowing one-click blocking of overseas users.
  3. Optimized the interstitial ad removal feature in Download Center, adding more algorithms to greatly improve recognition accuracy; ads are deleted only after being confirmed as ads, preventing accidental deletion of the main video.
  4. Optimized video upload: you can now monitor local computer folders directly in the backend and automatically scan and upload videos, subtitles, covers, and other files on a schedule.
  5. Optimized auto seeding: you can set auto seeding only for videos in specified categories, and instant segmentation mode also supports auto seeding.
  6. Optimized batch seeding: added full-series seeding; select any video associated with a series to seed the entire series with one click.
  7. Optimized system update: during system updates, you can view the complete update process log.
  1. TS inserted ads support position ranges, randomly placing ads within a specified interval.
  2. AI automatic editing supports field-specific output rules for titles, tags, actors and descriptions, such as asking AI to phrase content according to requirements or limit its length.
  3. Added subtitle keyword filtering with multiple rules. During transcoding, matching source watermarks, advertisements and translation-group credits are removed from subtitle files.
  4. Fixed playback failures when TS inserted ads lacked audio.
  1. AI editing and translation support the Google Gemini interface.
  2. Upload management supports direct episode uploads, automatically assigning completed videos to the corresponding series and sorting by episode number.
  3. Bidirectional data synchronization can skip TS and synchronize only M3U8 and images, consolidating M3U8 from multiple EFV instances into one for unified management.
  4. Improved color reproduction when transcoding Dolby Vision/HDR videos.
  5. Improved built-in CDN performance and added M3U8 playback statistics.
  1. Watermark/scrolling-text removal and addition can bind to multiple categories, applying only within those categories.
  2. All status-filter conditions in video, series, audio and album management can be combined for more precise filtering.
  3. Improved Mac cms data collection in the download center: collect only movies or series. Subsequent collections reimport series posters that previously failed due to network issues or were deleted.
  4. Batch-requeue selected failed downloads and prioritize selected items in the download center.
  5. Improved one-action episode addition in series and album management, recognizing more naming formats and sorting automatically.
  6. Improved bidirectional synchronization with more detailed logs and counts of some duplicate data during synchronization.
  7. Added scheduled cleanup of log files once they reach a specified size in supported environments, preventing excessive disk use.
  1. TS opening/mid-roll ads gain obfuscation: dynamically modify m3u8 so video and advertising segments have identical characteristics, greatly reducing detection and blocking by ad plugins. Details: View update.
  2. Website management can restrict a bound domain to api endpoints and m3u8, ts, png and other resources, returning 404 for other URLs.
  3. ai automatic filling can exclude fields such as category tags and descriptions; these fields remain untouched when ai fills content.
  4. Audio, video and series data retrieval can filter records already synchronized to cloud storage.
  5. Added automatic daily database backups, retaining the latest 7 days by default.
  6. Improved remote transcoding stability. When Vast creates a server, it prioritizes the closest location to the local server.
  7. Improved duplicate detection: similar titles with different punctuation or spacing also count as duplicates.
  8. Added system updates from administration: run the update script when a version is available without manually using SSH.
  1. Configure separate upload concurrency for posters, videos, subtitles and other files on the upload page.
  2. Added remote transcoding. The most resource-intensive stage runs remotely and results return automatically to the local server, fully supporting watermarks, subtitles, watermark removal and other functions. Use a self-hosted remote server or hourly vast.ai GPU servers. Create and delete vast servers as needed for temporary high-volume acceleration; administration can automatically create a specified GPU server and destroy it when processing finishes.
  3. Sort videos by daily playback counts in ascending or descending order.
  4. Improved series/album episode addition: select all search results and batch-add them in one action. Common naming formats are recognized, episode numbers extracted and episodes sorted automatically.
  5. Scan imports and multiuser FTP uploads automatically remove poster files after successful import.
  6. Improved bidirectional EFV synchronization to support senders with hotlink protection, segment ads and distribution routes enabled.
  7. Fast ass watermark replacement supports mid-roll segment ads and can be combined with opening/mid-roll advertising.
  8. Administration displays the version at bottom left and automatically announces new versions.
  1. Global opening/ending trimming supports category bindings and different trimming rules for one or more categories.
  2. Added AV1 transcoding. H.264 remains the default; enabling AV1 encodes to AV1 before slicing.
  3. Added AES image encryption with configurable AES type, KEY and IV.
  4. TS encryption supports custom KEY and IV, plus an endpoint to retrieve segment KEY and IV values.
  5. Added batch AI filling of descriptions, tags and other fields for selected categories or all videos/series.
  6. Added automatic AI filling of video descriptions and tags after transcoding.
  7. Improved AI settings with arbitrary AI endpoint URLs, API Key and model selection.
  8. Added adaptive portrait-video bitrate: reduce bitrate according to visual complexity while preserving the viewing experience and minimizing size.
  9. The download center supports base64-encoded txt image files, automatically decoding them into images after downloading.
  10. Added progress logs for cloud-storage synchronization, bidirectional synchronization, AI filling and translation, viewable directly in the interface.
  11. Bidirectional synchronization can target one or more categories.
  12. Server monitoring now displays GPU information and utilization.
  13. Added batch category creation and daily/monthly/total playback statistics.
  1. Global opening skipping also trims endings, automatically removing specified opening and ending durations during transcoding.
  2. Multibitrate transcoding adds 4K and original-resolution priority. When processing both original and other resolutions, options matching the original resolution are filtered out.
  3. File management adds batch deletion and batch addition of videos to the transcoding queue.
  4. Automatic episode arrangement in series and albums reports detailed missing-episode information for quick completion.
  5. Simplified watermark/scrolling-text settings. Moving image watermarks travel diagonally from bottom right to top left, pause briefly and move out.
  6. Fixed poster-plugin settings not taking effect in WordPress exports, leaving posters as WordPress media-library attachments.
  7. Bidirectional synchronization can remove duplicates on the sender: detected duplicates are filtered and deleted from the sending side.
  8. Reduced excessive memory use in the built-in CDN under high concurrency.
  9. Cloud synchronization can include published data only, allowing posters and metadata to be prepared before synchronization.
  1. Set a screenshot as poster in one action, applying a selected screenshot to poster 1 or poster 2 across N videos or every video in a selected category.
  2. Added quality enhancement for H.265/AV1 and other videos. When transcoding H.265/AV1, detect compression and intelligently compensate bitrate to preserve quality, with a configurable compensation ratio.
  3. Added TMDB metadata retrieval, filling video/series posters and details in over ten major languages, including English, Thai, German, Japanese, Korean, Vietnamese and French.
  4. Download-center endpoints now support individual and batch deletion alongside creation, update and query endpoints.
  5. Audio transcoding logs each processing step and error for troubleshooting.
  1. Added album management for audio collections such as albums, EP, single collections, audiobooks and podcasts, enabling unified classification and management.
  2. Added album import. Place all audio tracks, album details and posters on the server; import creates the album, then automatically adds sliced tracks and sorts them by episode number.
  3. Album management can query incomplete albums and fill missing tracks in one action, extracting correctly named audio records, adding them and sorting by episode number.
  4. Audio management adds batch M3U8 copying and renaming. Copy multiple M3U8 URLs and prepend, append, replace or remove strings in N audio titles.
  5. Audio-download requests can include album id and episode number. Once downloaded and sliced, the track is added to its album and sorted automatically.
  6. Added album retrieval api and merged the existing audio retrieval api, retrieving and filtering data from both album and audio management.
  7. Added album creation/update api: use api to create and update albums, including episodes and posters.
  8. Series and album management add m3u8 copying to retrieve every episode's m3u8 URLs in one action for other uses.
  1. Screenshot-composite posters adapt to landscape and portrait videos: landscape combines 4 screenshots in a 2x2 grid; portrait combines 2-4 horizontally.
  2. Improved poster uploads before or after video transcoding, automatically generating corresponding poster 1 and poster 2 after a successful upload.
  3. Added font selection for ASS/SRT subtitles, allowing preferred fonts when burning ASS, SRT or global ASS effects subtitles.
  4. The download center imports audio from S3 and OpenList. Improved series detection recognizes numeric filenames and Chinese naming forms such as episode N, installment N and part N, automatically adding downloaded videos to their series.
  5. Added server audio import for slicing in audio management, with automatic categorization.
  6. Added audio downloads, automatically imported into audio management for slicing after completion.
  7. Video uploads show disk usage and warn at 90%, preventing continued uploads from filling disks to 100% and disrupting operation.
  8. Improved hotlink allowlists. All batch video-management operations support category selection, processing every video in a specified category without selecting individual videos.
  1. Transcoding shows the estimated remaining time beside each resolution's progress bar, updated in real time.
  2. Watermark/scrolling-text removal supports template export/import to quickly share removal configurations across categories.
  3. Rewrote WordPress export with simpler settings and support for most WordPress themes and plugins.
  4. Added preferred subtitle-language extraction: when transcoding mkv and other videos with subtitle streams, extract and burn the selected language first.
  5. Improved batch renaming: prepend, append, replace or remove strings in N video titles.
  6. Opening/ending trimming can split at a specified timestamp and remove either the earlier or later portion.
  7. Fixed downloads becoming stuck due to network conditions on some URLs.
  1. Server imports and multiuser FTP uploads accept existing screenshots for use during transcoding and slicing.
  2. Improved timestamp protection and added strict download protection against video sniffing and downloading by tools or browsers.
  3. Improved automatic segment-ad removal in the download center, identifying and removing segment ads without manual rules.
  4. Added GPU driver status, warning when drivers are missing or abnormal for quick diagnosis.
  1. Improved preview m3u8 generation by combining ts segments from multiple random timestamps into the preview m3u8.
  2. Category-based server imports scan multiple levels of subfolders.
  3. Downloads support custom headers, including Referer and Cookie; the download api also accepts headers.
  4. With fast slicing enabled, H.264/H.265 video with non-AAC audio transcodes only the audio before fast slicing.
  5. Mac CMS collection adds slideshow images, screenshots, thumbnails, preview m3u8 and preview-video options.
  6. Improved video notifications: after automatic completion notification, manual publishing can notify again.
  7. Retranscoding adds TS duration filters to identify videos with excessively long TS segments for reprocessing.
  8. Improved response efficiency of some retrieval endpoints for large video datasets.