mold

Local AI image and video generation on your own GPU. Mold supports NVIDIA CUDA and Apple Silicon Metal, with a CLI, native desktop app, web studio, mobile companions, Discord bot, and REST/SSE API built on the same engine.

Documentation · Models · Desktop guide · API

The generation composer offers Edit prompt for spacious long-text editing, recent-prompt search and Clear with Undo. Native iOS and macOS Library Settings also let you hide date separators for a continuous grid.

Install

Stable release:

curl -fsSL https://raw.githubusercontent.com/utensils/mold/main/install.sh | sh

Nightly CLI from the latest published main build:

curl -fsSL https://raw.githubusercontent.com/utensils/mold/main/install.sh | MOLD_CHANNEL=nightly sh

Automatic cloud provisioning uses a verified container digest for the client’s exact source revision. Nightly/source builds wait for their matching container readiness manifest; an unpublished image fails before creating a paid host.

The installer selects a compatible build and verifies its checksum. Linux x86_64 clients without a visible NVIDIA GPU receive a GPU-free CLI. To use only remote GPU hosts, including from a machine with NVIDIA hardware:

curl -fsSL https://raw.githubusercontent.com/utensils/mold/main/install.sh | MOLD_BACKEND=cpu sh
mold host add                    # paste the link `mold host invite` printed on the GPU host
mold run "a cat" --host workstation

The CPU archive needs no NVIDIA driver or CUDA libraries. Local GPU generation requires a CUDA build (or Metal on macOS). See the installation guide for Nix, Arch, Windows, Android, and source builds. GH200, GB200, and GB300 require future linux/arm64 artifacts and are unsupported.

Mold no longer publishes new versions to crates.io. Existing registry versions are historical; use GitHub releases, Nix, Docker, AUR, or a source build for current versions.

Quick start

# Generate with the default model
mold run "a cat riding a motorcycle through neon-lit streets"

# Choose a model and reproducible seed
mold run flux-dev:q4 "a sunset over mountains" --seed 42

# Edit an image
mold run qwen-image-edit-2511:q4 "make the chair red" --image chair.png

# Edit from reference images (repeat --reference; order is semantic)
mold run flux2-klein:bf16 "the woman from image 1 wearing the glasses from image 2" \
  --reference person.jpg --reference glasses.jpg

# Qwen Image 2.1: up to 10 ordered references, a transparent cut-out, or 6-step turbo
mold run qwen-image-2.1 "put the jacket from image 1 on the person in image 2" \
  --image jacket.png --image person.jpg
mold run qwen-image-2.1 "a red paper lantern with a gold tassel" --transparent -o lantern.png
mold run qwen-image-2.1-turbo "a lighthouse on a basalt cliff at dusk, oil painting"

# Generate video
mold run ltx-video-0.9.6-distilled:bf16 "a fox in the snow" --frames 25

# Turn a photo into a 3D mesh
mold run hunyuan3d-mini-turbo --image chair.png -o chair.glb
mold run hunyuan3d-2mv-turbo --front chair-front.png --left chair-left.png --back chair-back.png -o chair.glb

# Upscale a Library video as a durable framewise job
mold video-upscale create clip.mp4 --wait

# Launch the web studio and API
mold serve

For API-only servers, set web_ui_enabled = false in config.toml (or MOLD_WEB_UI_ENABLED=false) and restart. The interface defaults on; disabling it leaves every API route available. NixOS: services.mold.webUi.enable = false;.

Models download automatically on first use. Generated media is saved locally with prompt, model, seed, and generation metadata. LoRA reuse tracks adapter file revisions as well as strengths; replacing an adapter invalidates cached weights. FLUX and SD3 retain at most 256 MiB of derived LoRA deltas per engine and release that cache when unloaded. Web, desktop and mobile reject empty video/audio attachments and files above 64 MiB before encoding, with inline errors that preserve the prior attachment. Mobile also enforces a 45 MiB combined media budget. Framewise video upscale also needs the host codec bridge: Nix packages and CUDA containers include it, while raw binary installs must provide ffmpeg and ffprobe on PATH before the server advertises that feature.

What it supports

• Models: FLUX.1, Flux.2, Stable Diffusion, Z-Image, Qwen-Image, Qwen Image 2.1, Wuerstchen, LTX Video, Wan, MiniMax H3, and Hunyuan3D. See the model catalog for variants and hardware requirements. Wan's 1.3B BF16 and 5B Q8/FP16 paths are performance-qualified on Apple Metal as well as CUDA; fp8-scaled Wan checkpoints remain CUDA-only. MiniMax H3's compact runtime is supported on SM89 CUDA and Apple Metal; Metal admission checks each request against live unified-memory headroom. FL2VA accepts text-only (no frame), first, last, or both boundary frames, with roughly 1,000 prompt tokens with paired frames and charging their additional memory. Distilled FL2VA tiers include Turbo 4/8-step, DMAD 4-step, PDMD 2/4-step, and FastVideo's FastH3 V2 8-step (CUDA-qualified). Queued FL2VA jobs preserve first/last endpoint roles through durable storage and replay, including all supported Turbo tiers. Forced-local H3 execution accepts one FL2VA request; batches, sequences, and Ref2VA reference uploads require the server route. Remote Ref2VA automatically selects authenticated uploads or bounded inline references on hosts with authentication disabled; no client API key is needed for the latter. Server license acceptance remains server-owned. Fresh H3 reference videos can use MOV/HEVC, WebM/VP9, MKV/AV1, ProRes and VFR MP4 when the authenticated host advertises video normalization. The host preserves already canonical clips or converts to H.264 MP4, then validates exact canonical facts before admission. This requires working server-side ffmpeg/ffprobe tools; the embedded Mac engine does not bundle them. Video-reference limits and compatibility. Qwen Image 2.1 (non-commercial Qwen Research License) edits from up to ten ordered references, renders transparent PNG/WebP cut-outs, runs native 2K presets, and ships BF16, INT8, FP8, GGUF and 6-step turbo tiers; its CUDA fast path is 2.5x faster at 1024² and 6x at 2K than v0.32.
• Images: text-to-image, image editing, inpainting, ControlNet, LoRA, identity photos, reference-image prompting, transparent backgrounds, prompt expansion, and upscaling, as PNG, JPEG or still WebP.
• Video and audio: text/image-to-video, sequences, clip continuation, lip dub, text-to-audio, and MP4 output with generated audio.
• 3D: single-image and named multiview-to-mesh with Hunyuan3D 2.0 and 2.1, automatic background removal, optional highlight and lighting removal, and Hunyuan3D Paint PBR materials in CUDA builds. Shape transformers can be locally derived as qualified FP8/GGUF tiers with mold quantize, which registers the derived tier on the host that made it. Results are published to the Library as binary glTF with a rendered poster tile, exportable as OBJ, an OBJ+PBR ZIP, STL, or PLY, or shared as a turntable GIF, APNG, or WebP. Durable text-to-3D, supplied-mesh texturing, and mesh-rebuild workflows remain available through the API/CLI. Apps generate 3-D objects directly from supplied images in the main generation screen. Painted prints expose their base-color, metallic-roughness, and normal maps as independent, digest-checked Library downloads on web, desktop, and mobile.
• Multiple machines: connect local, LAN, Tailscale, and RunPod hosts, then route work and browse one combined Library.
• Organization: title, favorite, tag, collect, restore, and manage prints across the desktop and web apps.

Curated models use the same short manifest title throughout the apps and mold list, such as FLUX.1 Dev Q4. The runnable ID (flux-dev:q4) remains available for commands, saved settings, and API requests. Model files keep their original names; third-party models keep their provider titles.

Model weights keep their own licenses. See each model page for terms and current platform support.

Mold Studio

Use View (native apps) or View original (web, desktop and mobile) beside an attached input to inspect it without replacing or removing it. References, source/target pictures, boundary frames, masks, control and identity pictures, and attached video/audio have separate viewing actions. Image previews fit the whole picture. Use Actual size on desktop and web, or pinch to zoom in the native iOS app, for close inspection. Done or Escape returns to the attachment or picker. Picker previews do not select or attach the picture. Queue viewing actions show the full submitted media, which may already have been fitted to the generation canvas; older machines without the original-input endpoint report it unavailable.

Every download, with requirements, is on the Download page.

Platform	Download
macOS	Mold Studio for Mac — built natively, Apple Silicon, macOS 26+
macOS	Mold Desktop for Mac — same app as Windows and Linux, macOS 12+
Windows	Mold Desktop for Windows — self-signed, see below
Linux	Mold Desktop for Linux — GPU-free, for remote GPU machines
iPhone and iPad	Mold Studio Companion (iOS 26) and Mold for iPhone (iOS 17) — TestFlight
Android	Mold for Android — signed APK, Android 7.0+

Two desktop apps on the Mac. Mold Studio is built natively for macOS: signed, notarized, self-updating through Sparkle, with mold's own Metal engine built in, and it talks to any mold serve on your network too. Generate, a merged multi-machine Library, Queue, Models and Machines, all in one native window. Library’s native machine and media filter menus keep their current selections visible. Both native apps keep neutral queue placeholders for jobs without source images; Mac restores fully local source attachments without stale archive warnings. Queue jobs open into grouped details and live previews; Discover starts with manifest models. Generate confirms submissions, offers explicit random or fixed seeds, and lets you resize the prompt editor and settings inspector. Native playback settings control autoplay/repeat, and oversized GUI input pictures are fitted automatically while library originals remain unchanged. The Generate controls stay at the bottom; the Library can create a collection from selected prints and hide collection members from general browsing across all their machine copies. The native iOS companion shares collection hiding and offers Add to Collection directly in the selection bar, including New Collection… to create and file selected prints together. Library machine filters scope collection counts and membership as well as trash, restore and permanent deletion to the selected machines, preserving copies elsewhere. Unavailable inventories and collections absent from a machine are distinguished. Native Mac and desktop Library sync let you choose a destination machine for Copy to or Sync All to, with an opt-in Keep in Sync choice per destination. Automatic runs follow gallery changes rather than a timer; Stop remains available and dismissed activity can be reopened. Copies are backups: deleting an original does not delete its copies. Sync owns only the organization changes it applied, preserves destination edits, and always propagates hidden source membership. Sync Details closes with Done, Return or Escape and successful reports omit issue controls. Intentionally removed local copies stay removed. Choose All Machines to act on every copy. Native Trash remains browsable for hidden collection members and offers Put Back, Delete Immediately and Empty Trash for the selected machines. Mold Studio for Mac guide

Download Mold Studio for Mac (Apple Silicon, macOS 26+)

It installs as Mold Studio.app. Mold Desktop, the same app that runs on Windows and Linux, installs as Mold.app and runs on macOS 12 or newer; the two sit side by side and connect to the same machines: Download Mold Desktop for Mac.

Mold Desktop is the desktop app on Windows and Linux too. It puts New image, Queue, My images, Styles, Machines, and Settings in one window, in plain words, with six themes, for local and remote generation, and pairs with the iPhone and Android companions. Explore the desktop app

Download Mold for Windows (x86_64) — a self-signed NSIS installer. The published build is CPU / remote-hosts only; see the desktop guide for the CUDA recipe. Verify and explicitly trust the release's mold-windows-self-signing.cert.cer before installing; the certificate is not publicly trusted and does not suppress SmartScreen on its own.

On Linux, the GPU-free desktop archive (mold-ai-desktop-bin on the AUR) drives remote GPU machines; for local CUDA use mold-ai-desktop from the AUR, nix build .#mold-desktop, or the devshell's desktop-build CUDA AppImage. See the desktop guide.

On iPhone and iPad, Mold Studio Companion is the native app beside Mold Studio for Mac: pair by scanning the Mac's code, generate on your machines, and follow renders from the Lock Screen. Generation options adapt to large text, and the full-screen viewer keeps its media actions accessible. Rotate an iPhone horizontally for landscape video across the display; rotate upright to restore gallery actions, or tap Close to return to the Library. Library New badges and unread counts are saved on each client independently. Media discovered after the initial machine baseline stays new during a Library visit. Viewing a print clears it immediately; leaving Library clears the eligible images shown during that visit. Hidden or filtered-out images and arrivals after departure stay new. Merged copies count once; hidden collections and Trash do not count. Existing read history is preserved on upgrade. Native icon badges follow the same history, with the existing permission or preference controls. It is in TestFlight while it is new. Mold Studio for iPhone & iPad guide

The native iOS companion uses concise render notifications with its app icon; tapping opens the finished print from the background or a cold launch, with notification activation completed on the main thread. Library long-press and drag previews retain their thumbnail loader, and source selection prefers a reachable machine copy; the notification icon keeps the same authored colors in light and dark appearances. Its Live Activity uses the system background material with matching semantic text, a full-width progress row, and a separate machine/queue footer. Generate Options marks the selected aspect with a checkmark; attaching the first/start frame chooses the closest supported aspect, and a closing frame preserves it. Options shows proportionate aspect icons, source-image fitting (centered Crop to fill by default), and an explicit Random seed default.

The native iOS companion offers Open Settings recovery for denied Photos, camera, nearby-network, notification and Live Activity access. Photos saving uses add-only authorization and safely writes pictures and videos; photo imports use the system picker. See the native permission audit.

The native iOS Library keeps its offline notice compact and offers scrollable details with every full machine name, preserving room for saved prints at large text sizes. Decorative Library badges stay within their tiles; the full machine names remain in accessibility labels and Info.

Native iOS Queue recovery actions use separate full-width rows; compact input thumbnails keep the model and prompt readable.

The native iOS Models screen offers visible per-model Unload and Unload All Models controls for the selected server; unloading keeps downloaded files.

Native iOS development: nix develop -c companion-dev watches Swift sources, builds and relaunches in Simulator. companion-run launches once; see the native iOS developer guide for simulator and build-directory overrides.

Android uses the same remote-only Mold Studio mobile interface. Download the signed universal APK directly; there is no zip to unpack:

Download the Android APK · nightly · Android installation guide

More ways to create

Preview generations directly in supported terminals:

mold run "a cat" --preview

Inline image generation in Ghostty with --preview

Run the engine where the GPU lives and connect from another machine:

mold serve  # authenticated GPU host; enable Remote Access or configure its relay
mold host invite                     # on the host: prints a one-use pairing link
mold host add                        # on the client: paste the link; saves the host by name
mold run "a cat" --host workstation  # or MOLD_HOST=workstation, or mold host use workstation
mold library download --latest --host workstation  # save that print's original file here

Use encrypted remote connections for LAN or access outside your network. Every client pins the host key from a private one-use invitation and receives its own revocable grant. The Lambda relay carries ciphertext; prompts, API credentials and media remain encrypted between client and host. Old proxy credentials require an upgrade and fresh pairing. Hosts may also opt into Regular direct HTTP access. Create a separate Regular invitation with mold host invite --regular --direct-url http://HOST:7680 --output /private/regular-invitation.json, then claim it with mold host add /private/regular-invitation.json --device-name 'My app'. Each app receives its own persistent device key; REST apps send it in X-API-Key. Keep the host operator key private and authentication enabled. See the persistent REST app-key guide for private profile storage and the Nix operator-key distinction.

--offload also applies to remote renders and durable sequences on GPU hosts. Add --no-save to keep one render out of a server's Library; the host still publishes the print and moves it straight to trash, so mold trash restore gets it back until retention sweeps it. It applies to renders a server performs — a local render has no Library, and refuses the flag rather than ignoring it.

Pass --fit beside --image when the source and the canvas disagree. Without it the picture decides the canvas; with it the canvas is what you asked for and the picture is resampled onto it — crop-fill trims the edges, pad-fit adds black borders, lanczos-resize stretches.

mold run flux-dev:q4 "a lighthouse at dusk" --image wide.png --fit crop-fill --width 1024 --height 1024

A sequence of several clips is scripted through the CLI/API. LTX-2 uses a latent motion tail; image-conditioned Wan and H3 FL2VA can continue from the last frame. Text-only Wan, legacy LTX-Video and H3 Ref2VA join independent clips. H3 sequences require an H3-enabled server. A complete script and job recovery workflow are in the video guide:

mold chain validate shot.toml
mold run --script shot.toml --output walk.mp4
mold jobs list
# After the previous part completes, keep its model recipe/canvas/fps:
mold run --script next.toml --continue-from PART1_JOB_ID --host plato --output part2.mp4

Each part accepts up to 16 stages. Linked parts can continue without an overall part-count or duration limit, each as a separate Library MP4. The new part snapshots its predecessor's tail and renders only new clips; it never rerenders or reencodes prior clips. Its first stage must be smooth with no authored first source conflict. Independent-clip families remain independent at the boundary.

A 3-D render can be one shot or a durable workflow. A workflow keeps every stage — the picture it starts from, its matted and delighted copies, the shape, the paint — as its own retained artifact, reports each stage as it changes state, and resumes after a restart. It lives on the machine that runs it:

mold mesh-workflow create --prompt "a small ceramic fox" --texture --follow
mold mesh-workflow create --mesh chair.glb --image chair-albedo.png
mold mesh-workflow list

Find weights to install, and watch them arrive:

mold search "anime style" --kind lora
mold pull flux-dev:q4
mold downloads watch

See the remote workflow and RunPod guides. Use mold queue to manage remote work and mold library to browse and organize the host's prints, including mold library download to save a print's original file, mold library source-media to recover the conditioning image a print was made from and mold trash delete to remove one permanently. Retained sources survive queue cleanup and remain available while any referencing library output exists, including in the trash. Matching retained input sets from newly completed jobs share storage while keeping per-job provenance. Native Mac and desktop library copies also retain those sources on supported destinations.

In the native Mac app, Use source is available in the print inspector and source viewer, including right-click menus. Choose a supported input role, or Choose model… to select an installed compatible model, recipe and machine without leaving the source. Successful reuse opens Generate; other draft settings are preserved where the selected recipe supports them.

Every GUI's print details lists retained source media, including reference images, opening/ending frames, masks, control images, audio and video. Source Media in the library browses those inputs with media-type, machine and text filters. Open a source to view its original and every use, including multiple roles or keyframe positions within the same print. Choose an input to use it in a new creation without restoring the original recipe. The native Mac and iOS Source Media sections include hidden collections, with the Library's collection, machine, media-type, sort, search and thumbnail-size controls. Other clients retain their existing collection visibility. This browser does not create independent uploads or pins: permanent deletion releases each owning print's retention, while other owners and Trash keep their sources. Unknown file types remain visible with an explanation when they cannot be attached. Identical verified source bytes appear once, with every owning print and input role still available. Bounded in-memory previews and incremental inventory loading make return visits faster without changing source retention. Older machines that do not provide verified content digests keep separate entries. Native source viewers identify legacy originals from their contents when opened and keep large-library access checks indexed so inspecting a source stays responsive.

Library source details expose the owning print/member address and optional content SHA-256, with original-media copy, save and share actions where supported. Use mold --host <machine> library source-media <filename> --member <unique-prefix> to download reusable original bytes. Generation input flags accept print:<filename> and source:<filename>#<member> on the selected host, alongside ordinary local paths. Prefix a literal local filename with ./ to escape these prefixes. Library-addressed inputs do not fall back to local inference.

Native Mac mirrors accept missing archive-only job IDs and generation durations, a short version matching the same version with a build suffix, and an alpha-channel fact derived from identical output bytes when absent on the source. The same compatibility check compares the gallery listing with its stable source archive, so completion bookkeeping does not falsely report a changed print. When an older archive omits scheduler or transparency settings, the source output is downloaded and its digest verified before its embedded recipe can corroborate those fields. Output bytes and all generation settings must still match; conflicting recorded provenance is refused. Retrying Sync All repairs retained inputs on compatible existing copies without duplicating their outputs.

Native Mac Save Locally and Sync All count a verified finished print as saved when its source explicitly reports legacy inputs as unavailable. The sync summary reports unavailable original inputs separately; it does not claim they were retained. Existing copies still require matching output bytes and recipes. Corrupt or unauthenticated input archives, changed outputs, and unsupported source transfer contracts remain failures.

Native Mac shipping builds include the reviewed Metal H3 runtime. Remote generation follows the selected host’s capabilities. Reused native Mac references show bounded previews and survive ordinary control edits and repeated submissions; relaunch verifies their byte-free origin locator before enabling Generate.

Native Mac local authoring inputs are saved privately alongside the prompt and scalar settings, including first/last frames and parked attachments. Unresolved retained references still require their original machine to validate them. Save failures are visible; missing or corrupt saved inputs block Generate until you reattach what you need and explicitly choose Use current inputs.

Before importing a source-bearing copy, clients check destination readiness. Windows local destinations currently cannot receive retained inputs, so those copies are refused before creating a local library output. Source-free copies remain supported, and Windows clients can recall retained sources from a supported remote machine. Unix storage-only destinations can receive retained-source backups when their gallery and encrypted media store are ready, even with generation workers disabled. To install Mold's Agent Skill for supported coding agents, run:

mold skill install --detected

For Claude Desktop and claude.ai, mold skill install claude-desktop writes the skill as a ZIP to upload under Customize > Skills and adds the local mold mcp connector to Desktop (guide).

The installed bundle uses each agent's native metadata and discovery contract, with a concise router, safety guidance, tested examples, and the prompting corpus: a shared guide, one complete base guide per manifest family (prompt style, syntax, generation context, examples, pitfalls, and that family's CLI examples), task leaves for the distinct H3, Wan, and LTX-2 grammars, and model leaves for checkpoints with quirks of their own. The corpus in crates/mold-core/src/prompting/ is also what mold expand, mold remix, --expand, the app Expand and Remix actions, and the MCP expand_prompt / remix_prompt tools hand to the LLM, together with the exact model, canvas, frame count, fps, and ordered references, so agents and the expander follow one set of rules. Hunyuan3D's base guide is the one that tells an agent NOT to write a prompt.

macOS Metal memory

Use mold system metal-memory status to inspect this Mac, or mold gpu list --json for a running host. Explicit root-only set <MiB> / reset commands support an optional boot policy with --persist. See the Metal memory guide for budget accounting, local-only administration and rollback semantics. Z-Image's Metal whole-decode path retries with tiles on memory errors and preserves the eager CPU fallback. CPU/CUDA decode ordering is unchanged; see the VAE qualification record.

Project

Mold is a Rust workspace built on candle. The documentation covers the CLI, configuration, deployment, and HTTP API.

Core contributors: James Brink and Jeffrey Dilley.

Licensed under the MIT License. Third-party components and model licenses are listed in THIRD_PARTY_NOTICES.md and the model documentation. InsightFace identity weights require separate acceptance and are limited to non-commercial research use.

Model checksums are verified when files are downloaded. Complete installed models queue and switch without full checksum scans, including after restart. To check existing bytes explicitly, run mold info MODEL --verify. Normal loading still checks file sizes and formats; it does not guarantee detection of same-size corruption.

The public website privacy policy describes the documentation site and Studio. Neither loads third-party analytics scripts; Studio keeps its encrypted pairing credentials in browser storage.

Native iOS and macOS Libraries offer All Media, Photos, Videos and 3D filters within each shelf. iOS Select supports finger sweeps across a range, including edge scrolling, and keeps the viewport stable when selecting.

Native macOS Settings ▸ Remote Access, beside Machines, offers Pair your phone in the existing first section, alongside the connection address and paired-device controls. Scan its QR in the native iOS app or Tauri mobile app. Saved machines advertise only routes their encrypted listener actually serves. Native clients prefer LAN and retain the same pinned host identity when moving to the Lambda WebSocket relay. Stopping Remote Access closes active sessions. Headless servers configure MOLD_PUBLIC_URL and MOLD_SECURE_WEBSOCKET_URL; the private engine remains on loopback. HTTPS Studio tries encrypted LAN first; allow local network access when the browser asks. Browsers that block local WebSockets use the encrypted WSS relay. Operator credentials stay on the host and are never supplied by the connector on behalf of a paired client.

Native iOS queue cards use concise curated model names and open full job details on tap, with state-aware per-job controls. Generate exposes searchable per-machine Prompt History with prompt-only recall. The companion resumes following accepted work after a temporary connection failure without submitting it again. Queue previews have bounded memory and concurrency, and both native apps require a definitive source lookup before reporting that a moved job's original was removed.

Native iOS keeps Cancel, Retry, and Pause/Resume in queue-row swipe actions, with Move to… on eligible queued, paused, and held cards and a read-only Details swipe when no recovery or pause action applies. Cancel remains available while source images and batch settings load; Retry appears when the current job supports recovery. Native iOS, macOS, web and Tauri job details expose the machine's saved failure diagnostic with copyable job identity. Errors explain known model, verification and graphics-device failures in plain English; technical details remain available for troubleshooting. Held cancellation rechecks the owning machine and job. Current Mold servers enforce held-only cancellation, preventing stale Held actions from stopping a render. Update older machines for this safeguard.

Native iOS and macOS generation controls support ordered MiniMax H3 image/video/audio references, Hunyuan3D named views, and Wan/MiniMax boundary frames. See the native reference parity audit and native app guides for limits and media formats.

For first/last-frame video authoring, the first frame controls automatic aspect matching when present; otherwise the last frame does. Removing the first frame or replacing a last-only frame updates the canvas to the remaining image's closest supported shape. Required-frame rules still come from the selected recipe.

Frame fitting defaults to Crop to Fill, including when only the last frame is attached. Supported fit choices apply to the submitted endpoints without replacing the original images in the editor.

Media export controls

Native iOS and Tauri offer clip GIF exports with Loop/Bounce, Forever/Once, size, frame rate and, on supporting hosts, an extra boundary pause. Set the pause to 0 ms to add no hold to a continuous source loop. GIF frame timing remains FPS-derived; a pause does not repair a discontinuity in the source. Native iOS also exports host-advertised mesh formats, turntables and texture sidecars to Share, Files or its Mold folder, with GIF delivery to Photos. See native export coverage.

Image inputs above a model’s processing size are prepared automatically. Native reference strips accept ordinary oversized pictures; processing pixel budgets are separate from bounded image ingestion and request transport limits. Large identity photographs are proportionally reduced to the face detector’s ingestion envelope, and original files remain untouched.

Queue rows show the owning machine’s sealed conditioning images across models. Job Details shows every ordered reference separately with its role, including identity photos, named views, masks, control images and boundary frames. Audio/video references are listed by kind; unavailable previews are disclosed. Inputs remain separate from live denoise previews and are fetched through authenticated routes, including work submitted from another device. Older servers retain their singular source preview.

Error messages

Errors use concise summaries instead of internal traces. Memory refusals show estimated requirements, the available budget and the shortfall in decimal GB/MB; a shared-memory Mac identifies the shared pool. These are estimates on the generation machine, not a promise that freeing exactly that amount guarantees a render. Full server diagnostics remain in logs and durable records. Both native apps share MoldClient error presentation, including older-server replies; local failures are recorded in OSLog with privacy-protected details.

Queue model downloads

A held job with missing downloadable model components offers Download and Retry on its queue card and in Job Details across native, web and Tauri clients. Current servers identify repairable missing companions, including adapters, even when the main model is installed. The original error, Failure Details and eligible Move to… control remain available during recovery. The job shows Starting, download-queue status, live bytes/progress, license review, reconnection and failures in place. Closing details does not stop recovery. The download runs on the job’s owning machine; retry occurs only after every returned download ticket succeeds and the original held job and server identity are revalidated. Failed or cancelled downloads leave the job held. Global queue pause stays in effect. Cancelling a job does not cancel a model download that other jobs may need.

Reuse input media

Reuse settings restores retained opening and closing frames into editable wells, preserving keyframe indices and ordered image references. Audio, source/continuation video, identity photos and control images restore with their saved settings where the selected recipe supports them. Continuation overlap and authored reference strength (including zero) are preserved. Older prints without recorded reference strength keep the server default. The original machine must retain the inputs; missing, damaged or oversized inputs are disclosed before generation. A new attachment or explicit removal wins over a late download, and restored wells never have a hidden archive fallback that can revive removed media.

Continuous Library rows

Library thumbnails on native macOS, native iOS, Tauri desktop and web use continuous justified rows: each picture keeps its aspect ratio, with narrow seams and square corners. Thumbnail size and window changes reflow the rows while keeping a visible print in place. The final incomplete row stays left aligned without enlarging a lone print. iOS thumbnails omit machine-name badges; image details and machine filtering still show the owning machines.

Deleting the media currently open in a Library viewer keeps the viewer open on the next surviving item in the current filtered order, or the previous item at the end. It returns to the grid only when no items remain. A surviving machine copy stays visible after a partial deletion failure. Native macOS successful moves to Trash finish without a persistent completion notice; progress, interrupted work and failures remain visible.

LTX convolutional video VAE decoding uses aspect-coupled spatial Auto tiles independently of temporal memory estimates. Long clips still require memory for temporal context and output accumulation; see spatial tiling.

All clients offer Secure (the default) and explicit Regular direct HTTP/HTTPS connections over LAN, Tailscale, or WAN. Enable MOLD_ALLOW_DIRECT_API=true on a headless host for Regular access; the managed proxy always requires end-to-end encrypted Secure transport. Connections never downgrade automatically. Regular HTTP relies on the network or VPN for encryption. Pairing again updates the same physical machine and its saved connection mode.

Secure clients make up to three connection attempts for transient proxy startup failures. Authentication failures stop immediately, and requests already sent are never automatically replayed.

Finding models

Model search in Generate, Installed and curated Discover accepts partial names, ordered abbreviations and small spelling mistakes. Case, accents and separators such as -, _, : and . do not matter. Try flx dev, flxu dev, or fluxdevq4; multiple words can match the model name, family or repository. Direct name matches rank ahead of fuzzy matches, and clearing search restores the existing list. Selection, machine routing and capability restrictions stay unchanged. Search is available even when only a few models are installed.

Add-on look selectors and supported upscaler/adapter selectors use the same matching behavior. Filtering a select keeps its current value visible.

Live Hugging Face/Civitai discovery uses the provider's normal search first. After a successful zero-result query, an updated server searches a broader prefix, filters at most 100 returned candidates with fuzzy matching, and paginates that matching set. Source, family, kind and mature-content filters still apply. Provider errors remain errors; this bounded fallback is not an exhaustive search of either provider's index. Older servers retain their provider search behavior.

Batch size applies to images and videos; choose 1 to make a single output. Native iOS exposes it in More Options. Cancel a whole batch from its header on native iOS/macOS, or open a queued job’s details and choose Cancel batch on web, desktop and Tauri mobile. The CLI equivalent is mold queue cancel --batch <BATCH-ID>. Completed outputs remain in the Library.

Native Library reliability: macOS viewer Delete moves the displayed print to Trash, selected thumbnails have an opaque outline and checkmark, and Library counts reuse revision-keyed projections during sync. iOS opened originals are cached as complete files for offline viewing and original sharing/saving; the first video open downloads the complete file before local playback. Native Mac and iOS storage preferences include Unlimited. Native macOS nightlies include bounded performance diagnostics; see the capture guide.

Native Source Media uses tight Library-style rows that preserve each original’s aspect ratio, with badges inside the image. Resizing and ordinary sync edits retain warm thumbnails. Source changes and Refresh revalidate inventories without timer-driven reloads; temporary thumbnail failures retry without scrolling.