Toggle "Place as karaoke captions" in Auto Caption to generate captions that highlight each word as it's spoken. Both preview and exported video include the karaoke effect. Pick from Classic / Pop / Glow / Bar styles and set a dedicated accent color per clip.
NewNew karaoke-captions mode in Auto Caption — each word lights up as it's spoken
NewFour karaoke style presets — Classic (sung words stay colored) / Pop (active char briefly enlarges) / Glow (accent-colored halo on the active char) / Bar (a progressive underline under sung words)
NewInspector exposes dedicated karaoke fields (accent color / style preset) — set once at generation time, fine-tune per clip
Generate captions automatically from any video or audio clip and drop them onto the timeline as text clips. Transcription happens entirely in your browser — audio never leaves the device. Japanese and English are supported, with optional translation of Japanese audio into English captions. See the auto-captions guide for details.
NewAuto-generate captions from any video or audio clip and place them on the timeline as text clips
NewChoose between Fast and High Quality models. Models are downloaded once into the browser cache and run offline thereafter
NewTranslate Japanese audio into English captions — pick "Translate to English" in the subtitle-language selector
NewLive preview of captions as inference progresses, inline text editing, and SRT / VTT / TXT downloads
Added a Cmd+K (Ctrl+K on Windows/Linux) command palette and a ?-key cheatsheet listing all keyboard shortcuts. Type what you want — like "cut", "split", or "undo" — and run it straight from the keyboard without hunting through menus.
NewCommand palette (Cmd+K / Ctrl+K) — type to run any timeline action, tool toggle, or playhead jump
NewKeyboard cheatsheet (`?` key) — every available shortcut grouped by category
Short animated clips (up to 30s) — fading-in titles, word-by-word kinetic typography, lower thirds — can now be added to the timeline by simply asking Claude in chat. Not intended to substitute for After Effects-level work; see the Claude-generated animated clips guide for capabilities and limits.
NewClaude can now build animated clips and place them on the timeline — prompts like "a 3-second opener with the title flying in from the left" produce a usable animated clip in place
NewPre-check mode for animation prompts — iterate on variations before any encoding, catching issues like text overflow or empty animations cheaply
Reworked the tool schemas around Anthropic's MCP best practices, making Claude operating VideoBuff noticeably faster end-to-end.
ImprovedRead-only tools now run in parallel — composite requests like "show me the project + history + diagnostics" resolve in one round-trip
NewBulk multi-clip edits — "apply this filter to all video clips" now happens in a single call
ImprovedSuppressed redundant post-edit verification calls, cutting response time ~50%
ImprovedTool definition tokens reduced ~30%, speeding up every Claude turn
ImprovedError responses now embed next-step hints so Claude self-recovers without retries
ImprovedClips identified by filename and time range, improving Claude's reasoning accuracy
FixedFixed connection conflict when running Claude Desktop and Claude Code together
Drive VideoBuff directly from Claude Desktop / Claude Code. Trim, add text, color grade — all from a chat message. See the Claude-integration guide for full walkthroughs.
NewAdded MCP integration so Claude can operate VideoBuff on your behalf
NewOne-click installable .mcpb bundle is now available
Videos with partial speed changes on split clips now export as a single seamless piece.
ImprovedContiguous clips from the same source are merged into one export unit
ImprovedFaster export for videos containing contiguous clips from the same source
FixedEliminated the faint click at split-clip boundaries
Automatic loudness matching for major platforms, plus reduced memory and improved stability for audio export.
NewAdded loudness (LUFS) presets to export settings (Web/Social -14, Apple/Podcast -16, Broadcast -23)
NewAdded tooltips explaining each export setting (resolution, codec, framerate, quality, loudness)
ImprovedDrastically reduced audio memory usage when exporting a short section trimmed from a long video
Improved video preview controls.
NewPlay/pause by clicking the preview area — same as pressing Space
ImprovedExtended zoom range to 5%–300%, making it easy to overview long videos
Significantly faster export and smoother preview playback.
ImprovedPipelined frame rendering and encoding for faster export
ImprovedDirect-to-disk writing via File System Access API drastically reduces memory usage
ImprovedParallel audio mixdown reduces processing time
ImprovedReduced unnecessary re-renders during preview playback for smoother performance
Long video export now possible. Significantly reduced memory usage and improved stability.
ImprovedOn-demand video data fetching enables export of long videos without running out of memory
Major improvements to clip drag & drop — intuitive placement for all clip types.
ImprovedClips now land exactly where you drop them
ImprovedDrop indicator shows landing position while dragging
ImprovedUnified drag behavior across all track types (video, audio, text, image)
ImprovedExtending image/text duration automatically pushes subsequent clips
Create gaps between video clips. Images from image tracks display as backgrounds in gaps.
NewGap placement between video clips
NewNumeric input for image/text display duration
VideoBuff — a browser-based video editor, now available.
NewMulti-track timeline
NewH.264 / H.265 export up to 4K
NewText & image overlays
NewColor grading (16 presets + 14 parameters)
New7 transition effects
NewSpeed ramping (0.1x to 10x)
NewAudio mixing (EQ, compressor, gate)
NewKeyboard shortcuts
NewProject persistence via OPFS