Version History

• Export presets for AI training datasets in JSONL, CSV, and txt-per-image formats
• Multi-caption sets for different training strategies
• Dataset health dashboard with quality and readiness metrics
• Caption and tag translation
• Token counting for common training architectures
• Expanded public API workflows and write operations

HEIC / iPhone Photo Support

• Upload iPhone HEIC photos in libraries, caption tools, and the public API.
• Improved iPhone HEIC uploads when a browser reports a generic or missing file type.
• Updated the HEIC decode fallback to recognize the installed decoder runtime.
• Decode full-resolution HEIC photos, including multi-tile grid images from iOS cameras.
• Accept valid iPhone HDR gain-map HEIC stills with bounded auxiliary image grids, while continuing to reject malformed, sequence, and non-still HEIF files.
• Show queued, completed, and failed deep-zoom processing states instead of leaving stalled images spinning.
• Restored automatic crop checks for compatible iPhone HEIC photos, including supported HDR gain-map stills.
• Improved library processing-card updates so completed media can recover from a missed realtime notification without requiring a manual refresh.
• Restored library thumbnails when a valid HEIC JPEG preview needs to be regenerated.
• Verify uploaded file contents against their declared type in both directions.

Media Editing and Review

• Improved image version comparison so zoom, right-drag pan, exit, and reopen actions remain reliable.
• Kept the follower track muted while synchronized audio versions are compared.
• Improved video timeline accessibility with consistent timecode labels, clearer zoom and fit status, and keyboard-operable IN and OUT handles.

Audio and Video Processing Reliability

• Improved direct audio uploads so transcription and speaker results settle reliably.
• Improved video-processing handoff so completed analysis remains available to metadata and review tools.
• Show generated video titles on library cards when processing completes.

Video Editing and Media Review

• Added video editing foundations for timeline navigation, scrubbing, zoom, transcript and caption overlays, frame markups, waveform views, and clip selection.
• Added caption rendering and export actions, including saves to new library media and selected-audio clip renders.
• Improved video thumbnails, representative frames, scene summaries, and processing retry behavior.
• Added Matroska video upload support and more reliable video artifact access.

Automatic Cropping and Image Editing

• Added an automatic crop workflow that suggests regions using object detection and lets users review, adjust, name, and queue crops.
• Added padding, aspect-ratio, thumbnail, focus, pan, and proposal controls for crop review.
• Added progress and fallback messages while automatic cropping runs or cannot produce a result.
• Improved duplicate detection, proposal persistence, coordinate handling, and stability across crop editing.

Media Libraries, Search, and Organization

• Added recoverable library trash so deleted media can be restored safely.
• Renamed departments to workspaces and improved search and filter discovery with tags and curated aspect-ratio options.
• Added visible media-processing progress, safer staged uploads, and more predictable Add More behavior.
• Improved upload admission, duplicate handling, and library thumbnail behavior across media types.

Browser Companion

• Added tag filtering, video thumbnails, capture and upload controls, and expanded media details.
• Improved connection, search, and live-site stability, including generic and hover-zoom captures.

Blind Assist, Public API, and Reliability

• Improved Blind Assist image results and narration playback, and restored reliable result delivery across the free tools.
• Introduced authenticated Public API v1 workflows for media, libraries, video, artifact downloads, transcripts, and collaboration data.
• Hardened account, organization, and private-media access while improving sign-in and redirect reliability.
• Improved PDF and audio library thumbnails, prompt validation, and processing status updates.

Media Libraries and Search

• Added department dashboards so teams can browse, search, preview, and open media across the libraries inside a department.
• Improved department and library search accuracy for exact text searches, short search terms, and cases where new media had not appeared in search yet.
• Added source-library labels in department-wide results so users can tell where each media item belongs.
• Added duplicate-upload protection so libraries can skip exact duplicate files and media revisions reject duplicate uploads.
• Added per-library custom icons and uploaded round image thumbnails for easier recognition in the collapsed sidebar.
• Fixed a production library crash that could happen when older media records were missing expected metadata.
• Strengthened upload validation so originals are checked more carefully before users rely on them in the library.

Audio and Export Improvements

• Added timestamp controls to audio transcript exports from library and department pages.
• Kept SRT, VTT, and structured transcript exports timestamped while letting plain text transcripts include or omit timestamps.
• Fixed transcript-only exports so a single requested transcript downloads directly in the chosen format instead of inside a ZIP.
• Removed blank metadata files from transcript-only exports when no metadata fields are selected.
• Updated export documentation and screenshots for the latest library export behavior.

Browser Companion Private Beta

• Prepared the Chrome and Edge Browser Companion for production-backed internal testing.
• Added workflows for searching Captionator media, downloading assets, inserting compatible media into web pages, uploading files, and capturing selected page media.
• Added Page PDF capture with a screenshot fallback for pages that cannot be converted cleanly.
• Improved Page PDF output for lazy-loaded images, nested scrolling areas, and wide page layouts.
• Added organization switching, permission-aware actions, and owner/admin session controls for Browser Companion access.
• Hardened live-site connection, upload, and library-refresh behavior for Browser Companion testing.

Permissions and Administration

• Added granular organization permissions with Simple and Granular modes.
• Added custom roles, department and library access grants, and reusable access groups.
• Improved Organization Security screens with better scrolling, clearer role editing, and constrained-window support.
• Fixed organization invite acceptance so returning or already-invited users can join reliably.
• Updated customer guides for organization permissions, invites, departments, libraries, chat, workflow setup, and media workflows.

Collaboration Polish

• Refreshed chat, group chat, and media chat screens with clearer layout, improved search, and better message handling.
• Improved the collaboration setup wizard, workflow editor, and library settings screens across desktop and smaller browser windows.
• Added and refreshed screenshot-backed user guides so teams can learn the current collaboration workflows from live-style examples.
• Improved notification display for media thumbnails, mentions, and dark-mode readability.

Processing and AI Platform Reliability

• Moved core processing workflows onto a more reliable background processing platform.
• Moved captioning, tag cleanup, document, audio, and vision workflows to dedicated processing services for more predictable results.
• Improved audio transcription reliability for noisy files and continued speaker-identification work.
• Improved PDF processing so completed documents can populate searchable titles, descriptions, and tags.
• Improved processing notifications so completed media updates refresh connected library views more reliably.

Platform and Site Quality

• Upgraded the application platform and deployment documentation for the current production environment.
• Improved CMS page rendering and public-site editing workflows.
• Fixed chat ordering so conversations display in the expected sequence.
• Improved backup and operations tooling for safer maintenance.

Team Collaboration

• Collaboration dashboard with personalized widgets: My Work, Activity Feed, and Team Overview
• Customizable review workflows with pre-built templates for common approval processes
• Kanban board for visual drag-and-drop workflow management
• Calendar view showing media items by due date
• Assign media to team members with due dates and progress tracking
• Real-time notifications for assignments, mentions, and status changes
• Built-in team chat with organization, department, and library channels
• User profiles with custom handles and profile photos
• Share media with external clients for review and approval
• Workflow automation rules for auto-assignment and status transitions
• Organization setup wizard for quick onboarding
• Role-based permissions: Owner, Admin, Manager, Member, and Viewer
• Activity timeline showing the full history of every media item
• In-app feedback form for submitting feature requests and suggestions

AI Content Detection

• Automatic scanning of images and videos for signs of AI generation
• C2PA and Content Credentials metadata reading
• Invisible watermark detection for popular AI generation tools
• Digital edit and manipulation analysis
• AI provenance filters to quickly find synthetic content in your library

Audio File Support

• Full audio file support: MP3, WAV, FLAC, OGG, and more
• Automatic speech-to-text transcription with word-level timestamps
• Speaker detection and custom speaker naming
• Audio editor with waveform visualization and spectrogram display
• Live captions synced to audio playback
• Audio trim with preview, save to library, or download
• Transcript export in TXT, SRT, VTT, and JSON formats
• Waveform thumbnails in the library grid
• Media type filters: Image, Audio, Video, and Document

Search and Quality

• Improved search performance for large libraries
• Search scope selector: search across all fields, titles, descriptions, or tags
• PDF first-page preview thumbnails
• Image cropping to create new images from existing media
• Automatic blur, nudity, and face detection
• Classification-based library filters
• Light and dark theme support

Detection and Filtering

• Library filters for blur, nudity, faces, and watermarks
• Detection confidence indicators in the media editor
• Faster library search for large collections

Platform Foundations

• Real-time media status updates as processing completes
• Live library refresh without manual page reload