A native Mac studio for
AI video, image, and audio

Generate with the models you choose. Edit, layer, transition, mix, and finish everything on a real Mac timeline.

Free download Bring your own provider keys No token subscription

0:00 / 0:00

Control the Work Around AI Production

Most AI video tools stop at the output. Studio keeps the entire production together: source assets, frame references, prompts, provider jobs, AI revisions, overlays, transitions, audio continuity, captions, exports, and every decision you may want to revisit.

Bring Your Own Providers

Route broad image and video generation through OpenRouter, create speech and sound with ElevenLabs, or work directly with MiniMax and Runway for video generation and editing. Your provider keys stay in Keychain, and the project remains yours when a model changes.

Lock First and Last Frames

Control continuity by pinning the opening and closing frames of a generation. Pull stills from existing assets, generated clips, or timeline segments and use them as frame locks, edit sources, or visual references. Studio keeps the inputs visible after you submit, so you can trace a result back to what produced it.

Assemble on a Real Timeline

Layer video on dedicated overlay tracks, add frame-accurate transitions, mix overlapping audio, and keep accidental visual collisions out of the edit. Trim, split, move, duplicate, and revise clips without leaving the project.

Audio, Voice, and Sound

Generate speech, sound effects, and ambience with ElevenLabs, design reusable voices, and transform recorded performances. Audio Continuity can turn separated dialogue clips into one reviewed performance while preserving intentional timeline timing.

Captions and Exports

Generate captions alongside your video, edit them inline, and export with burned-in subtitles or SRT/WebVTT sidecars. Timeline exports produce production-ready MP4s with your choice of caption treatment, and the export path is always a visible save panel, not a hidden cache folder.

One .studio Package

Every project is a single .studio package on your Mac. Generations, references, captions, exports, timeline edits, and a local SQLite project store live together in one folder. No cloud sync required, no disposable chat threads, and the project format is yours to keep, move, or version-control.

Model RouterChoose the right model
for the job

Use broad model access for generation, then route specialized video edits and audio work directly to the provider built for them.

38+Image models
  • GPT Image
  • Gemini
  • Seedream
  • FLUX
  • Krea
  • MAI-Image
  • Riverflow
  • Grok Imagine
  • Recraft
23+Video models
  • Seedance
  • Kling
  • Wan / HappyHorse
  • Veo
  • Hailuo / MiniMax
  • Sora
  • Runway
  • Grok Imagine
  • FLUX.3
Audio
ElevenLabs

Dialogue · Voiceovers · Sound Effects · Ambiance

Providers
  • OpenRouter
  • MiniMax H3
  • Runway
  • ElevenLabs

Set First & Last Frames

Pick the image a clip starts from and the image it should resolve to. Studio keeps the frame inputs beside the generated result, so continuity stays visible instead of disappearing into a prompt box.

Frame block
First frame showing a high-speed drone flight through a rocky canyon arch at sunset.
First frame
Last frame showing Earth from orbit with sunrise over the horizon.
Last frame
Generated video

Generate Your Video

Type what you want, generate the take, then review and place it on the timeline — without leaving the project.

Prompt
Main subject: young American woman, early 20s, everyday colonial appearance. Simple homespun linen gown in muted tan and faded indigo, rolled sleeves, work-worn apron, worn leather shoes. Brown hair loosely pinned up, soft face with light freckles, warm and grounded personality. Consistent identity throughout. Location: small Revolutionary War-era village, calm late morning. Dirt lanes, weathered clapboard houses, split-rail fences, stone well, kitchen gardens, clothesline, free-range chickens, mature oaks and elms, faint woodsmoke. No soldiers, battlefields, flags, or political symbols — just domestic life on the home front. Visual style: ultra-realistic historical realism, candid behavior, natural body language, unscripted slice-of-life, living-history documentary feel rather than costume drama. Camera style: intimate handheld vérité, as if recovered home footage. Handheld shake, imperfect framing, focus hunting, lens breathing, exposure shifts, motion blur, rolling shutter, mild compression artifacts, faded color, faint sensor noise. No stabilization. No cinematic moves. No stylized grading. 00:00–00:02 — Outside a clapboard house, she sits on a split-rail fence adjusting her hair, a contented smile crossing her face while focus drifts. 00:02–00:04 — Camera follows her into the side yard; she crouches to greet a barn cat, framing tilting as the operator stumbles. 00:04–00:06 — She scratches the cat behind the ears and offers it a scrap from her apron pocket. Autofocus pulses between her face and the animal. 00:06–00:08 — Beneath two broad trees, she lifts a wet linen sheet from a basket and drapes it over the clothesline, fabric billowing in the breeze. 00:08–00:10 — On the front-porch step with a pewter cup, she gazes over the dirt lane, brushing windblown hair from her eyes. 00:10–00:12 — Close side profile. Someone off-camera calls her name; she turns, smiles warmly, raises a hand, and says, "Morning." 00:12–00:15 — Walking down the dirt lane, pewter cup in hand, she notices the camera trailing her, gives a small half-smile, then looks away and keeps walking. Recording cuts abruptly to black. Audio: natural ambient sound only — songbirds, distant rooster, wind in trees, rustling leaves, cat's mew, footsteps on dirt, fabric on the line, creaking gate, faint voices, crackle of a cooking fire. No music. No sound design, no narration.
Output

Build the full soundtrack inside Studio

Generate voiceovers, dialogue, sound effects, and ambience with ElevenLabs. Revoice selected performances, lock approved words to their original timing, then mix, fade, duck, preview, and export from the same Studio project.

Keep every line in one continuous performance

Select connected speech clips and choose the result you need. Preserve Performance changes the voice while retaining the source delivery. Transcript Locked lets you approve the exact wording, then rebuilds each phrase against its original entrance and timing. Apply the replacement as one undoable edit.

  • Review exact wording
  • Phrase entrances preserved
  • Undo restores the originals

Voiceover and Dialogue

Create voiceovers, single-speaker dialogue, and two-person dialogue with compatible Voice Profiles. Review the exact request, generate the take, and place it on the Dialogue lane with its settings and provenance attached.

Sound Effects and Ambience

Turn a prompt into an effect or looping ambience. Choose automatic or precise 0.5–30 second timing, prompt influence, loop behavior, and output format before placing it on the Audio lane.

Voice Changer

Revoice a reviewed audio asset or timeline clip while preserving the original delivery. Studio extracts the chosen range locally, sends only that audio after approval, and imports the result as a non-destructive new take.

Voice Design

Describe a voice, generate auditions, and compare provider-written or custom preview scripts. Save only the take you choose as a reusable Voice Profile; discarded previews are cleaned up.

Timeline AI Editing

Edit the clip you already have

Select video on the timeline, describe what should change, and generate a reviewable replacement with MiniMax H3 or Runway. Studio keeps the timeline context and audio intact while you decide whether to use the result.

  • 01Start from the selected timeline video
  • 02Keep the existing timeline audio unchanged
  • 03Review the replacement before it enters the edit
Video 1 · Selected range Review first
Video 1
Instruction

Carry the warm sunset light through the selected shot while preserving the camera move.

MiniMax H3 Runway
New take Ready to review
Audio 1 unchanged

Everything lives in one .studio package

A Studio project is a folder on your Mac. Not a cloud workspace, not a disposable chat thread, and not a server-side blob. The package format is yours to keep, move, back up, or version-control.

Generations

Image, video, and audio outputs from OpenRouter, ElevenLabs, Runway, and MiniMax, stored with the prompt and settings that produced them.

References

Frame locks, edit sources, and visual references pinned to generations so you can trace a result back to its inputs.

Timeline

Multi-lane edits, video overlays, frame-accurate transitions, mix settings, trim points, and linked companion edits stored as package-local state.

Captions

Generated and edited captions, inline edits, and export preferences travel with the project.

Exports

MP4 exports and caption sidecars saved through a visible save panel, not a hidden cache folder.

SQLite Store

A local GRDB database tracks provider jobs, generation history, voice profiles, and assistant actions.

Ready to create with Studio?

Download free on the Mac App Store. Bring your own provider keys and start producing.

Get Studio on the App Store

Free download Bring your own provider keys No token subscription