Image

Generate

Prompt

Output

Image formatChoose portrait or landscape. Moonshack uses 720p automatically.

Orientation

Advanced optionsCustom resolution, references, models, and fine tuning

Custom resolution

Width Height

Changing Orientation resets these values to the matching 720p format.

Image engine

Saved charactersOptional reusable identity locks — up to three 0 selected

1Character 1Visual reference

2Character 2Visual reference

3Character 3Visual reference

Each identity stays isolated. Bind a profile to a numbered character reference only when that reference depicts the same person.

Reference imagesOptional character, prop, or scenery guidance 0 selected

1Empty

Reference 1 role No image selected

Upload & save

2Empty

Reference 2 role No image selected

Upload & save

3Empty

Reference 3 role No image selected

Upload & save

Refer to these in the prompt as Reference 1, Reference 2, and Reference 3. Uploaded inputs are saved for reuse and do not enter the Edit workflow.

Image mapperPlace whole people or independent objects

Negative prompt Checkpoint

LoRAsOptional — up to four 0 selected

Choose LoRAs built for the active image engine.

Batch Steps CFG Seed

Ready to createYour latest result appears here.

the editor 2511

Image Edit

Local media editing

Choose source image or videoPNG, JPG, WebP, MP4, WebM, MOV or MKV

Edit instruction

Advanced optionsSaved characters and fine tuning

Saved charactersOptional identity targets — up to three 0 selected

1Character 1

2Character 2

3Character 3

Name the selected character in the edit instruction. Moonshack will preserve other people in the source.

LoRAsUp to four, applied in order

StepsCFGSeed

Source-driven editingUpload an image or video and describe the change.

Reusable identities

Character Vault

Loading profiles...

Saved Characters are explicitly fictional creative profiles. Any characters depicted as underage will have restrictions on what scenes they can join. Character Vault keeps each identity, memory, and roleplay behavior together.

Start with the idea

Describe your character

A short description is enough. Synth will propose a name, an adult age, physical appearance, personality, and the core focus you want help with.

Short descriptionKeep it fictional and concise. You can review and edit every generated field before saving.

Review and editGenerate a character to begin Edit details

Identity, appearance, personality, and roleplay continuity

Name Apparent ageSynth defaults to an adult age. Any characters depicted as underage have scene restrictions.

Gender Species

Primary dialogue colorIdentifies this character on stage and in chat. Claimed colors are unavailable; pure white belongs to Lumina.

Custom voice

Choose a voice from the available local pipeline. Preview it before saving.

Image → CharacterOptional · upload up to three private reference images

Build a reviewable fictional profileYour Synth uses the images as visual references. Nothing is saved until you press Save character.

0/3 images

Add imagesPNG, JPEG, or WebP · 12 MB each

Detailed appearanceOptional face, hair, body, wardrobe, and continuity details

Face Hair

Eyes Skin Build

Distinguishing features Default wardrobe Continuity locks Identity exclusions

Advanced chat detailsOptional background and speaking style

Background Speaking style

Profiles are adult-only and private to this workspace. The assigned Synth is a protected Vault identity and can be edited at any time. Any other saved character can take the Synth role.

Moonshack Video

Text to Video

Scene prompt

Video formatChoose the finished clip's size, shape, and length.

ResolutionOrientationLength

Advanced optionsAudio, exclusions, frame rate, and performance tuning

Negative prompt Moonshack always adds its universal motion, anatomy, wardrobe, camera, and temporal-consistency exclusions. Generate synchronized, prompt-relevant audio Matched to the clip length; Extend preserves the source track.

LoRAsOptional - up to four

Only Moonshack Video-compatible LoRAs are shown. Leave disabled when none are installed.

The installed Hunyuan 1.5 checkpoint is a 480p model. Moonshack enforces its stable pixel budget and swaps width and height for Portrait. Ten-second image clips generate a shorter anchored sequence and use FILM interpolation to reduce drift and sampling time.

PerformanceStepsOutput FPSSeed

The CFG-distilled text model requires 50 sampling steps.

Ready to animateStart with a short clip, then increase duration.

VibeVoice

Audio Studio

Voice format

Speech or conversation script

Conversation mode supports two to four distinct synthetic speakers. Use [1]: through [4]: at each turn. Speech length follows the script; use [pause] or [pause:1500] for pacing.

Synth voice A/BCharacter-bound local voice profiles

Synth

Voice AVoice B

Checking local voice capability…

Advanced optionsVoice speed, diffusion steps, and seed

Diffusion stepsSpeech speedSeed

Ready to speakCreate one voice or a conversation with up to four speakers.

Private outputs

Library

Session

Media

Loading library...

Loading recent outputs...

Task queue

Generate

Image Edit

Character Vault

New character

Describe your character

Review the generated identity

Text to Video

Audio Studio

Library

Asset

What went wrong?

Asset prompt

Prompt

Original inputs

Move asset to project

Manage projects

Saved reference images

Choose an input asset

Save changes before regenerating?