AIMIX V1.4.9 Release: Conversational AI Editing, Talking-Head Cleanup and Long-to-Short

AIMIX V1.4.9 is here, and the headline feature is AI editing. The timeline editor gains a conversational AI panel: tell the AI in plain language to “remove the repetitions and pauses in this talking head”, “move the conclusion to the top”, or “cut three 15-second vertical shorts from this long video” — it handles speech recognition, proposes an edit plan, and applies it to the timeline after you preview and confirm. Long-to-short conversion, captions, music ducking, multi-timeline management, motion graphics and professional hand-off are all upgraded in this release.

What's new in this release

• Conversational AI editing: direct the AI in natural language right inside the timeline editor — preview plans, undo anything
• Edit by transcript: a word-level timestamped transcript where cutting sentences, reordering and fixing typos never touches the original footage
• Long video to shorts: one deliverable can combine non-contiguous segments; up to 30 clips per task, frame-accurate durations
• Multi-timeline management: create, duplicate, switch, rename timelines with independent aspect ratios
• Caption-card editing: split and merge cards, edit text, restyle in bulk, with bilingual caption tracks
• Automatic music ducking: background music dips automatically under speech
• Motion graphics: animated title and quote cards with transparent-channel export
• 19 built-in editing workflows: talking-head cleanup, long-to-short and other common cuts in one step
• Pro hand-off: export FCP XML and SRT for Premiere Pro and DaVinci Resolve
• International sign-in: email and Google account login

Conversational AI editing: tell it how to cut, and it cuts

V1.4.9 adds an AI editing panel on the left of the timeline editor. After you import footage, the AI first transcribes the speech into a word-level timestamped transcript — then you simply describe what you want, as you would to a human editor:

• "Remove the self-introduction and start with the substance"
• "That line is repeated — keep the second take"
• "Cut three 15-second vertical shorts, one theme each"

The AI turns your request into an edit plan that lists every cut and keep, and applies it in one pass after your confirmation. Everything supports undo, redo and named snapshots, so mistakes are always recoverable. Conversational editing can run on your own ChatGPT account (Codex); speech recognition consumes AIMIX credits.

Edit by transcript: cut talking heads like editing a document

The essence of talking-head editing is deciding what gets said. V1.4.9 turns the transcript into a directly executable editing surface:

• Cut the filler: delete by sentence or word; pauses, filler words and repetitions go in one sweep
• Reorder: move the closing conclusion to the top by simply reordering the transcript
• Reuse lines: the same sentence can appear twice in the final cut, independently
• Fix transcription errors: correct mis-recognized words without touching the timeline

The original footage is never modified — every change happens at the "what gets said" layer and can be reverted at any time.

Long video to shorts: freer clipping

The upgraded long-to-short feature lets one deliverable combine non-contiguous segments: highlights no longer need to be consecutive — the AI can gather the best moments scattered across the video into a single short. One task produces up to 30 clips, each with exact duration validation: ask for 15 seconds, get 15 seconds. Combined with multi-timeline management, versions for different platforms and aspect ratios live in the same project with separate preview and editing.

The AI has the full editing toolkit

This release opens the entire manual editor to the AI: split, trim, move, duplicate, speed changes, volume, fades, audio detachment, insert and ripple delete. AI and manual operations share the same timeline commands, so AI changes are visible immediately and you can keep fine-tuning by hand.

More importantly, the AI can see the result: it can pull source frames and the player composite to self-check — whether captions cover a face, whether a joint shows a black frame — and verifies before delivering.

Captions and music, refined

• Caption cards: split and merge cards, edit text and timing, restyle a whole section; transcript fixes and on-screen captions are independent, so editing captions never requires re-cutting audio
• Bilingual captions: add bilingual caption tracks to the same video for international audiences
• Music ducking: once the AI detects speech, background music dips automatically — clear voice, unobtrusive score

Motion graphics: put the frame in motion

Built-in programmable MG graphics — animated title cards, quote cards and emphasis effects — with adjustable parameters, local rendering, and transparent-channel export (WebM, ProRes 4444) that stays crisp over any footage.

Workflows and delivery

• 19 built-in workflows: talking-head cleanup, long-video repurposing and other common cuts standardized into repeatable flows the AI follows for consistent results
• Personal workflows: save your own editing habits as reusable workflows and apply them in one click
• Pro hand-off: export FCP XML projects and SRT captions for further polish in Premiere Pro or DaVinci Resolve
• Task-based export: failed renders retry only the failed items, finished files are kept, and every file is automatically checked for duration and integrity

Also in this release

• International edition supports email and Google account sign-in

AIMIX also covers the full content-marketing pipeline — batch editing, image and copy creation, and text extraction. Update to V1.4.9 and try conversational AI editing.

FAQ

What is AIMIX?

AIMIX is an AI content-marketing studio (desktop, Windows and macOS) covering batch editing, talking-head cleanup, long-to-short conversion, captions and voiceover, and image and copy creation — helping brands, merchants and creators produce marketing videos at scale.

How is conversational AI editing billed?

Conversational editing can run on your own ChatGPT account (Codex); speech recognition for talking heads consumes AIMIX credits. Exact usage is shown in the app.

Can I still adjust manually after AI editing?

Yes. Every AI change happens on the same timeline and can be fine-tuned manually afterwards, with undo, redo and named snapshots available at any time.

How do existing users upgrade to V1.4.9?

Installed users get the update automatically in the app; you can also grab the latest installer from the download page.