Changelog

New updates and improvements to Sovran.

Transcript preview in clip trimmer, looping clip playback, and video editor improvements

  • Transcript Preview While Trimming — The clip trimmer now shows the spoken words for your selected range as you adjust the start and end handles. Words that fall within your trim range are highlighted so you can cut precisely to the sentence or phrase without scrubbing back and forth.
  • Looping Clip Range Playback — When previewing or trimming a clip, the video player automatically loops within the selected time range. No more manually rewinding to re-check the same segment — it plays continuously until you move on.
  • Smoother Video Editor Playback — The video editor now premounts the next clip on the timeline before it starts, eliminating the brief stutter between back-to-back video clips during preview.

Auto Remix onboarding, guided spotlight tips, task launcher, and dashboard retention overview

  • Auto Remix Onboarding — New users can now go from zero to a finished ad in minutes. Upload any existing video, let Sovran auto-detect scenes, pick the clips you want, add background audio, and export — all in a guided four-step flow. Your clips and audio land in your project, ready to remix.
  • Guided Spotlight Tips — Context-aware tips that highlight specific UI elements at the right moment — like pointing out the AI analysis button the first time you open an asset, or surfacing the Ad Workshop after you approve your first clips.
  • Task Launcher — A quick-access panel on the dashboard surfaces your most relevant next actions: upload assets, run AI analysis, open the editor, or launch your first render — so you always know what to do next.
  • Dashboard Retention Overview — New charts on the dashboard show your team's activity trends — uploads, renders, and exports over time — making it easier to see where momentum is building and where things have stalled.

14 animated caption styles for the video editor

  • 14 Animated Caption Styles — The video editor now ships with 14 built-in animated caption styles: Blur, Bounce, Flicker, Glow, Highlighter, Karaoke, Pill, Rainbow, Scale, Spotlight, Typewriter, Typewriter Erase, Wave, and Word Highlight. Each style is word-level and fully customizable — mix and match font, color, and stroke settings on top of any animation.

Scene detection, clip remixing, Build in Editor, and auto captions with style customization

  • Scene Detection — Automatically split any video into individual scenes with sub-second precision. Scene cuts appear as clip candidates in the Create Clips tab, ready to approve, trim, or export.
  • Clip Remixing — Select 2-4 approved clips and remix them into every possible ordering. Sovran renders all permutations in the background so you can A/B test different story structures without manual editing.
  • Build in Editor — Select approved clips and send them straight to the Video Editor with one click. Each clip lands on the timeline pre-trimmed to its exact time range, so you can add captions, music, and text overlays immediately.
  • Auto Captions with Style Customization — Generate word-level captions for all clips on your timeline using AI transcription. Captions are automatically trimmed and aligned to each clip. Choose from caption templates like TikTok Native, Classic, or Minimal, then customize font size, colors, stroke weight, and highlight style.
  • Clip Preview — Click any clip candidate thumbnail to preview just that segment in a looping modal player, making it faster to review and approve clips without scrubbing through the full video.
  • Simplified Derived Asset View — Exported clips, remixes, and editor outputs now show a streamlined detail page focused on playback and reuse. Analysis, clipping, and trimming are reserved for original uploads where they add the most value.

AI Video Editor — edit your videos with natural language

  • AI Video Editor (Beta) — Chat with the video editor to add clips, music, voiceovers, text overlays, and more using natural language. Say things like "add my cooking clips" or "put a CTA at the end" and the AI finds your assets and places them on the timeline.
  • 14 AI Editing Tools — The AI assistant can search your assets, voiceovers, text library, and captions — then add, move, split, duplicate, reorder, or delete overlays. It can also change the canvas aspect ratio and background color.
  • Brand-Aware Suggestions — The AI automatically pulls in your Context Vault — product specs, brand guidelines, testimonials, and ad scripts — so its suggestions stay on-brand without extra prompting.
  • Smart Suggestion Chips — Context-aware quick actions that adapt to your timeline. Empty canvas? Get starter suggestions. Already have clips? Get suggestions for music, CTAs, or format changes.

Create Clips from any video, new Asset Detail page, and upload reliability improvements

  • Create Clips — Turn any long-form video into short, ready-to-use clips. Sovran uses AI visual analysis to identify scene boundaries and suggest clip candidates automatically. Review the suggestions, pick the ones you want, and export them as new assets in one click.
  • Asset Detail Page — Click any asset to open a dedicated detail page with a full-size player, transcript, clips, and an edit view — all in one place. Quickly jump between tabs to review, trim, or create clips without leaving the page.
  • Visual Analysis — Assets now get automatic visual analysis that categorizes scenes by content type — product shots, lifestyle footage, testimonials, and more. Use the results to filter and find the right footage faster.
  • Asset Origin Tracking — Every asset now shows where it came from — uploaded, AI-generated, clipped from another video, or synced from Meta. Filter your library by origin to quickly find what you need.
  • Upload Reliability — Rebuilt the media upload pipeline with better codec detection, real-time status tracking with automatic polling fallback, and improved error recovery. Uploads are now more reliable across all video formats.

Video editor undo/redo, saved text overlays, TikTok Sans font, and rotated video fixes

  • Timeline Undo & Redo — The video editor timeline now supports full undo and redo. Every edit — trimming, reordering, adding or removing layers — can be reversed or reapplied instantly.
  • Saved Text Overlays — Save your favorite text overlays for quick reuse across projects. Saved overlays are cached locally for instant access when you open the editor.
  • TikTok Sans Font — TikTok Sans is now available as a font option in the video editor and text overlays. Includes all weight variants from regular to bold.
  • Rotated Video Support — Videos shot in portrait mode on phones now render correctly. Sovran detects rotation metadata and applies the right transform so your renders always look as intended.
  • Improved Media Browsing — The audio and video library panels in the editor now support pagination and search, making it faster to find the right asset in large projects.
  • Consistent Font Rendering — Font families are now normalized across the editor preview and final render output, eliminating mismatches between what you see and what you get.

Connect your Meta ad account and get weekly AI-powered creative suggestions

  • Weekly AI Suggestions — Connect your Meta ad account in the Context Vault and every Monday we'll analyze your top-performing ads, rank them by Ad Score (results relative to cost), and generate fresh creative suggestions — text hooks, visual hook directions, and voiceover scripts based on what's already working. For image ads, the AI sees your actual creative to give more specific feedback. Suggestions land in a new tab in your Messaging Library where you can accept the ones you like and save them for your next ad. You'll get an email each week when new suggestions are ready.

Team workspaces, improved signup, smarter AI, and reliability fixes

  • Team Workspaces — Collaborate with your team on video ads. Invite teammates by email, assign roles (Owner, Admin, Member), and manage permissions from the new Team page. Owners have full control, Admins can invite and remove members, and Members can view and edit all projects.
  • Claude Sonnet 4.6 for AI Generation — Script generation and text overlay variations now use Claude Sonnet 4.6, Anthropic's latest model — producing higher-quality ad copy, hooks, and CTAs.
  • Audio Rendering Fixes — Fixed audio stuttering and sync issues in rendered video ads. The audio merge pipeline has been rebuilt to handle edge cases that previously caused glitchy playback.
  • Smoother Video Uploads — Uploaded videos are now automatically normalized on your device before upload, fixing issues caused by variable frame rates and non-standard encodings from screen recorders and phones.
  • Video Download Reliability — Downloads are faster and more reliable, with proper Unicode filename support. ZIP exports now try the CDN directly before falling back to proxy.
  • Background Render Submissions — You can now minimize the render modal while your video ad is being submitted — no need to wait on the submission screen.
  • Free Marketing Tools — New free Creative Testing Calculator and Creative Testing Matrix tools to help you plan and scale your ad testing strategy.

Kling 3 — a new AI video generation model to choose from

  • Kling 3 Video Generation — Select Kling 3 alongside Veo 3.1 and Sora 2 when generating AI video clips. Kling 3 supports text-to-video and image-to-video modes with configurable duration and aspect ratio.

Ad Workshop — a redesigned sequencer built around how you actually think about modular ads

Ad Workshop — a redesigned sequencer built around how you actually think about modular ads
  • Timeline-Style Block Layout — Blocks now flow left-to-right like a video timeline. Each block shows clip thumbnails inline, and you can drag-and-drop to reorder.
  • Framework Templates — Start new sequences from proven structures like Hook → Body → CTA, Classic Direct Response, or Hook Swap Test. Pick a template and your blocks are pre-populated.
  • Live Combination Counter — A formula bar below the blocks shows exactly how many ad variations your sequence will produce (e.g., 3 hooks × 2 bodies × 1 CTA = 6 variations).
  • Upload & Create Directly in Blocks — Upload clips or create text variations from each block's “+ Add” menu. Generated variants are automatically added to the block and tagged with its label.
  • Custom Block Names — Name blocks whatever fits your creative process. Use the preset suggestions or type your own — blocks are just numbered slots until you rename them.
  • Simplified Navigation — The sidebar is now Strategy, Asset Bank, Ad Workshop, and Launched Ads. The Ad Workshop is accessible immediately — no need to upload assets first.

Full-featured Video Editor with Timeline, Asset Browser, and Auto-Captions

  • Video Editor — A new timeline-based video editor for creating custom video content. Drag and drop clips, add text overlays, and arrange your assets on a multi-track timeline.
  • Built-in Asset Browser — Browse your project's video library directly in the editor sidebar. Search for indexed videos using semantic search. Also includes access to Pexels stock video and image library.
  • Voiceover with Auto-Captions — Add generated voiceovers from your project to the timeline. Captions are automatically synced and added when you drop a voiceover onto the audio track.
  • Project Audio Integration — Access all your uploaded music files and generated voiceovers in the Audio tab. Stock music is also available alongside your project files.
  • Asset Pre-loading — Select assets on the Assets page and click "Edit Video" to open the editor with those assets already on the timeline, ready to edit.
  • Cloud Rendering — Export your video projects with cloud-based rendering. Progress updates in real-time and the finished video appears in your asset library.

Messaging Library with Personas, Angles, and AI Generation

  • Messaging Library — A new project-scoped library to save and reuse text overlays and voiceover scripts. Search, filter, favorite, and organize your messaging across all your video ads.
  • Personas & Angles — Define target audience personas with pain points, outcomes, and tone hints. Create messaging angles with do/don't say guidelines. Use them to generate targeted ad copy.
  • AI-Powered Generation — Generate text overlays and voiceover scripts with AI using your personas, angles, and Context Vault as input. Save results directly to your library.
  • Library Integrations — Insert text from your library directly in the Batch Text Overlay modal, Asset Editor, and Voiceover modal. Usage is tracked automatically.

Asset management and UI stability fixes

  • Asset Download Fix — Fixed an issue where downloading assets would open in a new tab instead of triggering a download. Downloads now work reliably across all browsers.
  • Asset Thumbnail Display — Fixed a race condition that caused video thumbnails to be missing after upload. Thumbnails now appear immediately after uploading assets.
  • Sidebar Collapsed State — Fixed the Active Project dropdown in the sidebar to display correctly when the sidebar is minimized.

Veo 3.1 AI Video Generation Upgrade

  • Veo 3.1 Upgrade — Upgraded to Google's latest Veo 3.1 models with higher-fidelity video output and improved motion quality.
  • Context-Aware Audio — Veo 3.1 generates smarter audio that better matches the visual content of your videos.
  • Last Frame Interpolation — New feature to create smooth transitions between a starting image and ending image in your AI-generated videos.

AI Voiceovers and Voiceover Stitch Videos

  • AI Voiceover Generation — Generate professional voiceovers from text scripts using AI voices. Preview voices before generating and create multiple voiceover variations.
  • Voiceover Stitch Videos — Combine AI voiceovers with your video clips to create new video assets. Select multiple videos to stitch together with your voiceover audio.
  • Auto-Captioning for Voiceovers — Voiceover stitch videos automatically include captions synced to the voiceover audio, making your videos accessible and engaging.

B-Roll Remix and Text Variations

  • B-Roll Remix — Select 2-4 video clips and automatically generate all permutations with optional text overlays. Perfect for testing different hook-body-CTA orderings.
  • Text Variations — New 3-step wizard to batch apply text overlays across multiple videos. Add multiple text variants to generate all combinations at once.

AI-powered video analysis, streamlined onboarding, and more

  • AI Video Analysis — Automatically analyze and classify video content with AI-powered module detection for hooks, testimonials, and more.
  • Hook-Body Onboarding Flow — New streamlined onboarding that guides you through creating your first video ad with hook and body clips.
  • Asset Editing & Trimming — Trim video clips directly in the app and batch edit multiple assets with captions and overlays.
  • Caption Auto-Generation — Automatically generate captions from video transcripts with inline editing support.
  • Meta Ads Integration — Configure launch defaults and publish video ads directly to Meta Ads Manager.
  • Various UI Improvements — Various UI improvements and bug fixes.