Before you
call action.

moviefactory.ai connects your film through Movie Brain, the production graph and continuity: story context, versioned production history and reviewed footage evidence. This first-of-its-kind connected filmmaking workflow brings planning, direction, editing and delivery to your browser and a dedicated native macOS app. Create a 3D version for VR or spatial viewing by importing the exported movie into Immersive Converter for Mac. You direct the story and decide what enters the final cut.

The studio keeps the whole film connected as you create. Movie Brain carries story context into each shot, the production graph keeps references and take history, and continuity tools help you review what belongs together. You can refine one part, explore another ending or compare takes while keeping earlier work. That connected workflow supports ambitious stories, with the filmmaker making the final decisions.

Movie Brain connects your screenplay, approved references and shot direction to reviewed frame and video candidates in the browser and native Mac app. Each operation draws on relevant story excerpts, pinned cast and world references, scene conditions, camera intent and neighboring shots. The actual inputs stay attached to the resulting candidate. Browser director plans use reviewed pinned context; native Local Director records the current scene, style and assigned library references. Inspect what informed a result and decide whether it belongs in the film.

The production graph connects reference versions, shot specifications, source media and approved cuts. Trace a candidate, review changes and keep earlier work. A new take retains the inputs and parent that produced it. Browser reference review exposes affected shots and lets you choose which adopt a changed version. Mac reference review can update the selected scene or all scenes. Older takes keep their original history. Generation creates a candidate, and explicit acceptance records the decision to use it in the cut.

Pin reviewed cast and world references to your shots, with explicit wardrobe, prop and scene conditions. When a reference changes, select the shots that should adopt it. Use the continuity inspector and footage index to compare intended state with timestamped frames and reviewed observations. Matching reference versions describe the plan; the footage still needs your review for identity, appearance and performance.

Movie Brain assembles context, the production graph records the versions and decisions behind each take, and continuity brings actual footage evidence into review. You can rehearse a camera move, generate a shot with pinned references, adjust a take locally, compare its change report and approve a candidate without losing earlier work. The production workbench brings footage search, alternative continuations and camera rehearsal into the same project.

Yes. moviefactory.ai includes a dedicated native macOS app with a first-of-its-kind connected filmmaking workflow. Write and plan, build reusable assets and 3D sets, approve references, generate tracked frames and compare numbered takes. Native tools render titles and logos, shape timeline motion, measure and finish audio, link originals with playback proxies, and deliver up to 4K in H.264, HEVC or ProRes. Save local documents and .mfarchive backups; use supported production packages or .movieassets packs for browser exchange. Native motion, transitions, normalization and ducking are preserved as data but are not rendered by the browser. Transfers are explicit, without automatic synchronization.

The native app requires macOS 26 or later. Core editing, importing, saving and local movie rendering work without provider credentials. AI generation needs your own eligible provider account or compatible configured workflow, and provider charges may apply. Joining the early-access waitlist registers your interest; invitations are shared separately and signup does not immediately unlock the app.

Write or import text and Fountain, save screenplay versions and split the story into scenes. Review AI writing and production-plan drafts before applying them. Discuss the film with the Production Assistant using text or optional voice. Title and lower-third directions become timed on-screen text, separate from picture prompts. Find these controls in Script → Versions, Import, Split by Scene Headings. Optional voice conversation requires microphone permission and on-device speech recognition, sends transcripts to OpenAI and stops when you close the assistant or choose Stop Voice.

Approve the film’s reference versions and shot intent, then update individual shot reviews as needed. Generate tracked opening and closing frames as candidates for Use Result; review a new opening’s references before generating the closing. OpenAI image settings offer three model choices, seven sizes, up to sixteen references and high-quality JPEG or lossless PNG. Find these controls in Shot Frames → Review References… → Approve Pinned Context. A finished request does not choose its result for you. Prepared tracked ComfyUI batch export remains a browser workflow.

An activity banner above the workspace shows the current operation, animated progress and elapsed time. Reference and shot-frame panels also show running, ready, candidate-review or error states beside the relevant image controls, and prevent duplicate pending requests. These states remain visible when you change rooms. Image generation has no promised percentage or completion time. Stopping after submission does not mean the provider stopped; check Generation Activity for an uncertain request before trying again. Images that cannot fit the film’s media budget remain available in Generation Activity → Recovered Files.

Compare synchronized players with stable take numbers, remembered source starts and speeds, and one audible side. Keep alternates, reject takes, save notes and move through the review queue. Importing a candidate leaves the cut intact until you choose Use in Cut. Find these controls in Generate Video → Take Review. Shorter sources hold their final frame during comparison. Selecting a take for the cut is explicit and undoable.

Block a camera move with local 3D stand-ins and export its storyboard or scene. Search footage, inspect continuity evidence and prepare alternative continuations. Local Director adjusts exposure, contrast and crop, or reduces source-clip volume, with multi-shot plans and fixed painted grading regions. Review candidates before acceptance. Find these controls in Header → Production Tools · Generate Video → Local Director. Fixed painted regions do not track moving objects or generate replacements. Camera geometry is not automatically sent to video generators, and native Local Director clip gain reduces volume rather than amplifying it.

Trim on the 24 fps editing grid, snap to cuts, audio and the playhead, and optionally ripple downstream audio. The Timeline menu moves a start or end by one frame with Undo. Set source trim and speed, add zoom, position and opacity keys, fades or a dissolve from the previous picture. Find these controls in Timeline menu · Timeline → Motion. The editing grid stays at 24 fps even when you choose another delivery frame rate. Studio and timeline playback render the assembled cut with its graphics and sound; the selected-scene preview has separate playback controls.

Import or generate speech, music and effects; trim layers, set volume and fades, and inspect decoded waveforms and source peak/RMS levels. Measure the mix’s integrated loudness and sample peaks, apply undoable normalization, and duck music beneath speech layers. The finished mix carries into WAV and movie delivery. Find these controls in Audio → Mix Finishing → Measure Mix Loudness. Native delivery supports up to 64 audio layers. Peak readings are sample peaks, not oversampled true peaks.

Link large source files, relink the same recorded media after a move, and create lightweight 540p playback proxies. Preview can use a proxy; delivery rechecks and reads the original. A native portable archive collects the files and verifies them when the film is restored. Find these controls in Assets → Link Original File… · Relink Verified Original… · Create 540p Proxy. Keep linked originals available. Relinking accepts the recorded media bytes at another location; different footage belongs in a new take. The proxy is a disposable playback cache and is not used as the delivery original.

Export a Final Cut bundle with FCPXML, editable cut ranges of the rendered picture master and a separate stereo WAV mix. Titles, speed changes and picture finishing are baked into that master. This handoff does not include source handles, editable Motion titles or separate audio stems. Find these controls in Edit → Final Cut Delivery….

Connect OpenAI, Runway, Higgsfield, ElevenLabs or your ComfyUI server, with credentials stored in macOS Keychain and sharing enabled per provider. Higgsfield offers Veo 3.1 and Fast at 720p/1080p with eight-second shots, plus Seedance 2.5 at 720p with a fifteen-second shot limit. Core editing, saving and local rendering work without provider credentials; generation depends on your account or installed workflow. Find these controls in Settings → Providers · Production Assistant. ComfyUI needs a running server, compatible installed models and nodes, and an exported API workflow with the required input mappings. Sharing consent is specific to the provider; credentials stay out of project exports.

Save editable work as a .moviefactory document; embedded media stays inside it and linked originals stay outside. A native .mfarchive collects originals, excluded shots and production history, and verifies files when restored on Mac. Export Portable Backup creates a supported filmlab-v1 JSON production package for browser or iOS exchange, with separate media limits and excluded scenes left out. A .movieassets pack carries a selected reusable asset version and its required dependencies. Transfers are explicit, without automatic device synchronization. Native motion, transitions, normalization and ducking survive compatible JSON transfers as data, but the browser does not render those effects. Export a movie when the finished picture and sound need to travel.

Independent filmmakers, first-time storytellers and anyone with a story to tell. Start with your own screenplay, footage or a guided template, then build one scene at a time. Advanced shot, camera and editing controls are there when you need them. Provider access and available workflows depend on the tools you connect.

moviefactory.ai brings planning, generation and editing into one workspace, so you can develop a film from your own creative direction. Plan with storyboards and camera rehearsal, combine your own footage with generated shots, and refine existing takes. AI generation uses your connected provider account and may consume credits, or runs through your own compatible hardware. Your costs depend on the project, tools and settings you choose.

moviefactory.ai is where you make the film: develop, direct, edit and export it. moviefactory.io helps films reach audiences, with promotion and monetization opportunities subject to film review and available commercial terms. Making a film in moviefactory.ai does not automatically publish it on moviefactory.io. Submission review and any commercial arrangements are handled separately.

Review the shot’s pinned references and scene intent before generation. The studio records the source inputs and prepared media sent to the selected provider, then saves the output as a candidate with its history. Accepting a result checks that the source context and candidate still match the reviewed work. A completed render does not replace the selected frame or approved take on its own. Reference slots, sound and other controls depend on the configured provider.

Yes. Scoped reference adoption shows the previous and proposed versions, the selected shots and their intended state. Approve that scope to append new shot specifications while retaining earlier versions, takes and timeline history. A costume change for one scene can stay in that scene. The review checks for changes made after the plan was prepared.

The local director handles exposure, contrast, warmth, same-aspect crop and reframe, trim and clip audio gain on supported media. Review the operation, source interval and preservation limits before rendering. Each edit creates a candidate alongside its source. Browser and native Mac processing use their own supported codecs and media limits; the plan identifies the available operation. A crop reframes the existing image, while physical camera changes require a new shot or supported generation workflow.

Yes. Use supported instructions or explicit controls to build a batch across selected shots, then review every operation before processing. You can combine warmth, contrast, exposure, crop, trim and clip audio changes. Ambiguous or unsupported requests need revision. Candidates are saved together only after the batch finishes; compare each result and accept the complete batch into the cut. Stale source inputs or failed processing block the batch.

The report separates recorded information, such as the operation, references and source history, from measurements taken from actual local media. Request a comparison to inspect aligned timestamped frames, duration and measured brightness or color differences. An unknown value stays unknown. Pixel measurements help you inspect a change; conclusions about character identity, expression or performance remain review decisions.

Index an actual clip to capture timestamped sample frames and audio measurements. Add observations at those moments, or explicitly share sampled frames with a configured provider for captions and visible text, then confirm or reject the suggestions. Search opens the matching footage at its timestamp and identifies the evidence source and review status. Sampling can miss an action; it does not inspect every frame or transcribe speech. Script and prompt descriptions remain separate from observed footage.

Yes. Choose an opening interval to keep and prepare two independent continuation branches from its boundary frame. Import ending clips or review generation through a supported first-and-last-frame provider. Compose each candidate locally, watch the transition and explicitly accept the ending you prefer. Earlier media and the alternative remain available. Composition re-encodes the movie, so matching file bytes or identical encoded pixels are not promised.

Add your email to the early-access list to help shape a studio for independent creators. We’ll share news about moviefactory.ai beta access. Signing up registers your interest; invitations will be shared separately.

Build full films, cinematic scenes, storyboards, animatics, trailers and concept films. Start from a story or screenplay, or upload your own images, clips and sound to assemble an edit.

A template starts with what you are making. Five layers describe the production type, genre, format, visual style and delivery requirements. The website’s composer previews how those choices can propose a brief, story structure, reference needs, shot plan and timeline across Script → Cast → World → Direct → Motion → Edit. Guided studio templates provide starting timelines and production planning keeps the brief, references and review checkpoints connected. Factual formats call for source review and clear labelling of generated reconstructions.

Yes. The moviefactory.ai companion app for iPhone and iPad records scene by scene against a shooting plan in portrait, landscape or square, lets you select takes, and prepares a production package with the story and the chosen clips. Import it in the studio’s project chooser and it opens in Edit & export as its own project, keeping trims, playback speed and framing. Nothing is uploaded to a server during the handoff.

The eight launch templates are Narrative short film, Vertical microdrama series, Mini-documentary, Educational explainer, Commercial / brand film, UGC product story, Music video and Interview / podcast. Their guided five-beat timelines are built into the studio, with genre, visual style and delivery packs applied when you create a project. The website previews the proposed briefs, reference needs, review checkpoints and delivery plans around those starting points.

Yes. Paste your screenplay into the Script room or use Import script to upload a .fountain or .txt file up to 2 MB. Analyze the script to propose characters, locations, environments, props and shots, then review the breakdown before applying it. You can generate the character and world references when you apply the plan, or create them individually later.

Yes. Keep characters, complete sets, locations, props and vehicles in a library shared across your films on the same device. Add reference views, named variants, tags, collections, favorites and source notes. Capture existing film references or camera blockouts, or import your own images and GLB models. Approve a version before assigning it to selected shots. The browser studio and native Mac app both support the library.

Each film receives an independent copy of the approved asset version, selected variant and its files. Editing or deleting the library entry leaves those film copies intact. Complete sets pin exact approved versions of their members, so a later costume or prop revision does not silently change an existing set. Reusing an asset updates the selected shot references; it does not replace footage or approve a new take.

Export a .movieassets pack containing a selected asset version, its original files and all assets required by its complete set. Import the pack into the browser studio or native Mac app. Export each version you want to retain: a pack does not automatically contain the entire catalog or its history. A film backup preserves only the asset copies adopted into that film. Transfer is explicit and does not automatically synchronize devices.

Start with references proposed from your script, add your own images, or reuse an approved asset from your library. Refine a character’s appearance or describe a location, environment or prop, then generate reference images through your connected provider. Keep different views and costume variants together. Assign the selected version to shots and review the actual generated footage for consistency.

You can start with your script and follow the production steps: review the breakdown, create your cast and locations, prepare shots, generate takes and assemble the edit. Refine one scene at a time, with advanced camera, lens and style controls available when you want them.

Built-in connections include OpenAI for writing and GPT Image 2 or 2.5; Runway and Higgsfield for Veo 3.1 and Veo 3.1 Fast, and Higgsfield for Seedance 2.5; ElevenLabs for speech and sound effects; and ComfyUI for custom generation. ComfyUI presets support MiniMax H3 turbo and LTX 2.5 distilled when their required models and nodes are installed. Other models such as FLUX, Qwen-Image and Wan need a compatible custom ComfyUI workflow.

No. Built-in connections need an eligible provider account and credentials. ComfyUI presets require the matching models and nodes on your server. Custom-workflow entries are examples of ComfyUI pipelines you can configure, not preinstalled or individually verified integrations. Model access, cost, hardware and workflow compatibility vary.

Generation uses the connected provider account and its credits, or your own ComfyUI hardware. Costs and available models depend on the provider and settings you choose. Provider API access and credits may be separate from consumer subscriptions.

Yes. Upload your own media, reuse images as shot endpoints, combine still frames with video clips, and select a soundtrack. You can build an animatic with stills before generating motion.

The workflow is Instruction → Reviewed plan → Execution → Candidate take → Filmmaker approval. For a request such as “Make the second shot darker. Keep the timing and sound,” the plan identifies the selected shot, operation, preservation limits, source media and processing destination. Local edits and reviewed multi-shot batches keep source takes intact. Provider work shows cost when known. The studio checks for stale plans, and a candidate enters the cut only when you accept it.

Retakes are scoped to a selected take, with a time interval and optional painted region for exposure, contrast or warmth changes. Set feathering and review the edges as the subject moves: the mask stays fixed in the image. Crop, trim and clip sound have their own scope. Every result is a separate candidate and the original media stays available. These local tools do not track objects or generate replacements. MP4 re-encoding can change encoded pixels outside the adjustment. Compare the result before accepting it into the cut.

Camera rehearsal lets you arrange actor and prop stand-ins, save camera positions and targets, play a path through the set and export PNG storyboards or an animated GLB scene. Use this local blockout to review composition before a shoot or generation request. You can also save a rehearsal as a reusable set asset. Geometry and camera paths are not automatically supplied to video generators.

Yes. Import self-contained GLB models with textures, rigs and animation into the asset library. Place independent instances in a complete set, select approved variants, adjust position, rotation, scale, materials, lighting and camera, and preview animation clips. Export a storyboard PNG or a composed GLB retaining supported geometry, textures, rigs and animation. This works in the browser and Mac app. It is a composition tool, not a mesh sculptor or rig authoring tool; imported geometry is not automatically used as AI generation conditioning.

Yes. In the browser, switch on CC or Show subtitles in Standard or Pro Edit to preview dialogue from the screenplay against the current cut. Export SRT or WebVTT files and a readable transcript, or choose subtitle burn-in separately during movie export. Timing is estimated from the screenplay and needs review against the finished film; this is not automatic speech transcription. The preview toggle does not change your export settings.

No. Browser projects and uploaded media stay on the device, and the Mac app saves local film documents. Configured AI operations may use a provider or a local rendering worker. The shared capability layer records each integration’s supported inputs, limits, availability, license record and test status. Local worker compatibility must be validated on the actual hardware. Each execution identifies assets, destination and purpose before external processing, with no silent cloud fallback or automatic transfer of an entire private project.

Production controls include persistent queued, running, completed, failed and cancelled job records; reconciliation of uncertain submissions; attempt and concurrency limits; and approved spending boundaries. Plans show execution location and estimates when available, with an unknown cost kept unknown. Permissions, protected credentials, asset provenance, component-license checks and no retention of client media for training by default are part of the platform. Existing provider accounts keep their own terms and charges.

Browser projects and uploaded media are saved on this device. Mac saves local .moviefactory documents with embedded media or links to original files; linked originals must remain available. A native .mfarchive collects originals and production history for Mac backup. Use a supported production package for browser or iOS exchange, with its separate media limits. The iPhone and iPad companion app hands off a shooting plan and selected clips. Tools and pending-edit acceptance rules vary by platform, and transfer does not imply automatic cloud synchronization. Browser provider keys stay in the session and are excluded from project exports.

Yes. Import alternative MP4, WebM or QuickTime takes in the browser, up to 50 MB per clip. Filenames can suggest shot assignments; confirm each destination before import. Compare takes in synchronized A/B views or filmstrips, add notes, mark Keep or Reject, and explicitly choose the take to use. Importing alternatives does not automatically replace the approved cut.

Yes. Add timed titles, lower thirds and end cards in the browser or Mac app, with shared typography, colors, placement and a logo. Preview them over the cut and render them into the movie. The browser also offers a clean export without on-screen text and accepts PNG, JPEG or WebP logos. Mac Titles & Graphics renders the shared text layers in native preview and movie export; import logos into the local film before using them.

Arrange dialogue, music and effects on separate audio layers, trim their sources, place cues, set fades, mute layers and adjust levels. Bring your own recordings or use configured speech, effects and compatible music tools. Mac adds measured loudness, RMS and sample peaks, undoable normalization and speech-driven music ducking. Export a stereo WAV mix. Browser WAV exports contain the audio layers; Mac WAV exports also include enabled source-clip sound. Movie exports include enabled clip sound as well as the soundtrack. Native normalization and ducking render on Mac; peak readings are sample peaks, not oversampled true peaks.

Use your OpenAI connection to translate screenplay subtitles, with sixteen language choices on Mac, then review meaning, names and timing. The browser also translates on-screen text. Browser and Mac workflows export SRT, WebVTT, transcripts and scene-based chapter lists; publishable chapter lists need suitable scene lengths. Subtitle timing is estimated from the screenplay and cut, not transcribed from audio. Translation uses your provider account and may incur charges.

Yes. Save a full-resolution PNG still from a readable source in the browser or Mac app. Review the chosen moment before saving. For a 3D set or camera rehearsal, export a storyboard PNG directly from its preview.

Choose GPT Image 2 or GPT Image 2.5 Sunburst or Flare, image size and quality. High quality can use JPEG or lossless PNG, and shot frames follow the film’s orientation. The Mac app supports up to 16 OpenAI reference images and prepares upload copies while keeping originals. Returned Mac images that exceed the film’s media budget remain available in Generation Activity → Recovered Files. Image generation requires your configured provider access.

Yes. Assemble your clips, still frames and soundtrack, then export a movie in the browser. Encoding method and export time depend on the browser, hardware and project; follow the export screen’s guidance. Choose 720p or 1080p MP4, landscape, portrait or square framing, and crop-to-fill or fit-with-bars. Browser exports support films up to ten minutes. Mac renders cuts up to two hours at 720p, 1080p or 2160p in H.264 or HEVC MP4, or ProRes 422 MOV, with eight frame rates from 23.976 to 60 fps and an SDR, 8-bit picture pipeline. Its Final Cut bundle contains editable cut ranges of a rendered picture master and a separate stereo WAV mix; graphics and finishing are baked in, without source handles, editable Motion titles or audio stems. You can download the screenplay, an edit decision list and a portable production package. Delivery uses approved timeline versions, landscape, vertical and square presets with reviewed reframing, and validated adapters for compatible editing workflows. Review the finished export and keep a restorable project backup.

Yes. Export your finished movie from moviefactory.ai, then import the file into Immersive Converter, a separate Mac app. Preview and adjust the depth, then export side-by-side or top-bottom stereo for compatible VR headset players, including those on Meta Quest. MV-HEVC spatial video supports Apple Vision Pro playback when exported on a compatible Mac. This is framed stereoscopic movie viewing; conversion does not automatically create 180°/360° video or a scene you can walk around. It remains an export-and-import workflow, with the converter’s own requirements and availability.