VELVET VICE LTX — Prompt Director

The brain of the Velvet Vice workflow — the node that actually writes your LTX prompt

By Velvet-Vice·Created 25 days ago·Updated 10 days ago· 2

image

prompt_package
final_prompt_preview
status

◄modeMANUAL►

◄manual_prompt►

◄short_idea►

◄full_auto_settingsMODE: FULL_AUTO DURATION: AUTO — READ FROM LTX DIRECTOR DURATION CHOREOGRAPHY: AUTO — SAFE DEVELOPMENT BEATS SCENE COMPLEXITY: AUTO — DURATION BUDGET MODIFIER PARTICIPANT REGISTRY: AUTO — PRESERVE 1 TO 4 VISIBLE ADULTS ANATOMY CLASSIFICATION: WOMAN / FUTANARI / OTHER / UNCLEAR BEFORE ACTION CONTACT STATE MACHINE: ON GEOMETRY COMPOSER: LITERAL POSTURE + SUPPORT + PAIRWISE RELATIONS LEGACY FALLBACK: PRESERVE EXISTING PROVEN ACTION WHEN UNCERTAIN SCENE STYLE: passionate, explicit, physically natural and continuity-focused INTENSITY: PASSIONATE ACTION COMMITMENT: HIGH RHYTHM: STRONG CONTROLLED ESCALATION SOFT TISSUE PHYSICS: REALISTIC INERTIA: BODY-DRIVEN DAMPING: NATURAL RESPONSE SCALE: MOTION-MATCHED OSCILLATION LIMIT: SINGLE DAMPED SETTLE CONTACT DEFORMATION: ON CONTACT ANCHOR: SURFACE-LOCKED VOLUME PRESERVATION: ON GRAVITY RESPONSE: ON ENDING: CONTROLLED BY THE ENDING MODE SELECTOR CAMERA: PRESERVE VISIBLE NUDITY AUTO-TRIGGER: ON DIALOGUE: none unless clearly requested AUDIO: natural breathing, vocal reactions and environmental sound only; no music PERFORMANCE STYLE: LIFELIKE AND REACTIVE BODY LANGUAGE: WEIGHT-AWARE, ASYMMETRIC, CAUSAL FACIAL PERFORMANCE: CONTEXTUAL MICRO-REACTIONS MOTION TIMING: ORGANIC, NON-METRONOMIC AUDIO SYNC: ACTION-BOUND DIEGETIC SUPPORTING MOTION BUDGET: 1-2 RELEVANT CUES PER BEAT POSITION RECOGNITION: AUTO-DETECT FROM VISIBLE GEOMETRY POSITION MODE: PRESERVE CURRENT WHEN ACTIVE POSITION TRANSITION: MAXIMUM ONE LOW-COST TRANSITION ROLE ASSIGNMENT: GEOMETRY-BASED, ANATOMY-OWNER LOCKED MIXED ANATOMY RECOGNITION: STRICT LITERAL VISUAL EVIDENCE FUTANARI ANATOMY LOCK: PRESERVE BREASTS + ATTACHED PENIS PER OWNER PERSISTENT ANATOMY EXISTENCE: PRESENT_LOCKED SURVIVES OCCLUSION + RECEIVING ROLE OCCLUDED HAND COMPLETION: FORBIDDEN — PRESERVE VISIBLE SILHOUETTE ACTION CLASSIFICATION: SEPARATE EFFECTOR + TARGET FROM POSITION ACTION LOCK: PRESERVE MANUAL / ORAL / PENETRATIVE / OTHER CLASS POSITION VARIANT: AUTO-DETECT INDEPENDENTLY CONCURRENT ACTIONS: AUTO — ONE PRIMARY + ONE COMPATIBLE SECONDARY GENITAL REINTERPRETATION: FORBIDDEN The user has activated the separate graph-level confirmation that every depicted participant is 18 or older and consenting. The supported configuration may contain one to four adults already visible in the reference image with any visible adult anatomy or presentation. Treat the image as the literal first frame. Assign stable participants A through D as needed from visible screen position and identity cues; preserve their faces, hair, clothing, limbs, anatomy ownership, depth and support points. First resolve anatomy ownership, then detect the current position from independent geometry: each posture, facing direction, relative height, front/back order, pelvis relationship, support points and existing contact graph. Preserve an already active position instead of replacing it with a familiar alternative. If the position is neutral or uncertain, keep the visible postures and choose only the smallest physically reachable continuation. Assign initiator, receiver, active anatomy and target anatomy from existing contact, explicit request, visible compatibility, shortest continuous path, stable support and camera visibility—not from assumed gender. This includes solo scenes and any supported one-to-four-adult combination, including participants with similar or mixed visible anatomy. A feminine-presenting adult with breasts and a naturally attached visible penis is a valid futanari anatomy configuration: preserve the breasts and penis together, keep the penis attached to its original owner, and never replace it with a vulva-only body. For two or more futanari participants, resolve and lock each visible participant independently so every confirmed visible penis remains present. Treat anatomy existence separately from visibility: temporary body overlap, camera cropping, anal/vaginal receiving roles or an occluded pelvis never remove or relabel a `PRESENT_LOCKED` penis. A partially visible or occluded hand remains partial or occluded; never complete hidden fingers, add a sixth digit, duplicate a thumb or invent a new hand path. Never exchange, duplicate, erase, normalize or reinterpret anatomy. Independently classify the current or requested action from its active effector and exact target: hand-driven, mouth-driven, pelvis/anatomy-driven, body contact, visible object, preparation or unclear. A mounted/rider/cowgirl label describes position only and must never overwrite the action class. Preserve a confident manual, oral or other action exactly instead of converting it to a stereotypical alternative. Select one physically achievable primary adult action that best continues the image, with at most one compatible secondary action after resource validation. Use at most one low-cost position transition and only when duration and visible supports make every step achievable. Keep all existing image-fidelity, lifelike performance, physics, Ending Control and action-bound audio rules active.►

◄adult_confirmedfalse►

◄ollama_modelfredrezones55/Qwen3.5-Uncensored-HauhauCS-Aggressive:9b►

◄ollama_urlhttp://127.0.0.1:11434►

◄ollama_context_profile8-12 GB►

◄ending_modeAUTO►

◄ltx_prompt_profileLTX 2.3►

LTX is a picky generator. Feed it a short prompt and it goes static or chaotic; feed it long specific prose and it rewards you. That's why the biggest node in the Velvet Vice pack isn't a sampler or an encoder - it's a prompt factory. VelvetViceLTXPromptDirector takes your raw idea and produces the fully structured prompt that actually drives the render, with four input paths and separate profiles for LTX 2.3 and LTX 2.5.

Four modes, two philosophies

The mode dropdown is the whole decision tree:

MANUAL - your manual_prompt passes through as exact text, untouched. Crucially, Manual mode never contacts Ollama. No LLM, no server, no risk. This is your escape hatch and your baseline.
STANDARD VISION - sends your short idea plus the reference image to a local LLM via Ollama for a grounded rewrite.
ADULT ASSISTED / ADULT FULL AUTO - the pack's adult-oriented modes. They use a much larger, uncensored instruction set that covers participant/anatomy ownership, geometry and contact planning, and action-bound audio guidance. The adult_confirmed toggle is the pack's explicit 18+ consent gate, and full_auto_settings is a large structured settings document steering the LLM. It's an adult pack, so if that's not what you're here for, MANUAL and STANDARD VISION do everything else.

Everything about its design fights the two classic LLM-enhancer failure modes the community hits - dirty output and subject drift: the profile enforces strict image-fidelity rules, the ending control block is a hard override appended to the prompt, and the final text routes through the Final Prompt Preview node so you can read it before committing a render.

The LTX 2.3 vs 2.5 profiles

The optional ltx_prompt_profile selector defaults to LTX 2.3. The 2.5 profile changes the prompt's DNA: it treats the connected image as the exact first frame, emphasizes motion and chronological action instead of re-describing the source, keeps a continuous take, integrates caused audio, and prepares a grounded brief for the workflow's optional Gemma prompt enhancer (TextGenerateLTX2Prompt). Manual prompts stay exact passthrough in both profiles. If the widget gets omitted by ComfyUI, the pack resolves the profile from workflow metadata, then from LTX 2.5 model markers, and falls back to 2.3 - so old workflow files don't silently change behavior.

The inputs that matter

mode - manual vs LLM paths, above.
manual_prompt / short_idea - your text. One is the exact output, the other the seed for the LLM.
ollama_model (default fredrezones55/Qwen3.5-Uncensored-HauhauCS-Aggressive:9b) and ollama_url (http://127.0.0.1:11434) - which local model, where. You can point it at any model you've pulled; this default is the pack's pick.
ollama_context_profile - 8-12 GB / 16 GB / 24-32+ GB - sets the LLM context window to match your RAM; pick too big for your system and you've found the classic self-inflicted OOM.
ending_mode - AUTO / NO CLIMAX / CLIMAX / LOOP. In adult modes it's a hard override on the ending's content.
adult_confirmed - the age/consent gate for the adult modes.

Outputs

prompt_package - a custom VELVET_VICE_PROMPT_PACKAGE for the Ollama Release Barrier, which forces the LLM out of memory before LTX loads.
final_prompt_preview - a STRING of the final prompt, for the Final Prompt Preview node.
status - a STRING with run diagnostics.

Installing it

The pack is VELVET VICE - LTX (velvet-vice-ltx) via ComfyUI Manager or:

cd ComfyUI/custom_nodes
git clone https://github.com/Velvet-Vice/velvet-vice-ltx

Restart, hard-refresh with Ctrl+F5. Zero pip deps, but the non-manual modes require a running Ollama (ollama serve) at the configured URL, and the LTX 2.5 branch needs the Gemma text encoder (text_encoders/LTX 2.5/gemma4_e2b_it_bf16.safetensors) plus the LTX 2.3/2.5 model set from the Civitai release.

Troubleshooting

LLM modes fail instantly - Ollama isn't running or the model isn't pulled. ollama pull <model> then retry; or switch to MANUAL, which never touches Ollama.
OOM while prompting - your context profile is too big for your RAM; drop to 8-12 GB.
Prompt ignored your idea - subject drift; read it in Final Prompt Preview, tighten short_idea, or run MANUAL.
Workflow behavior changed after an update - a stale selector or old pack version resolved the wrong profile; reinstall clean rather than merging folders.

It's what makes the workflow feel like a director instead of a prompt box - and showing the final text before the render is the honest place to catch what it got wrong.

CategoryVELVET VICE/LTX

Inputs (11)

Name	Type	Default	Description
mode	COMBO	MANUAL	4 options: MANUAL, STANDARD VISION, ADULT ASSISTED, ADULT FULL AUTO
manual_prompt	STRING		—
short_idea	STRING		—
full_auto_settings	STRING	MODE: FULL_AUTO DURATION: AUTO — READ FROM LTX DIRECTOR DURATION CHOREOGRAPHY: AUTO — SAFE DEVELOPMENT BEATS SCENE COMPLEXITY: AUTO — DURATION BUDGET MODIFIER PARTICIPANT REGISTRY: AUTO — PRESERVE 1 TO 4 VISIBLE ADULTS ANATOMY CLASSIFICATION: WOMAN / FUTANARI / OTHER / UNCLEAR BEFORE ACTION CONTACT STATE MACHINE: ON GEOMETRY COMPOSER: LITERAL POSTURE + SUPPORT + PAIRWISE RELATIONS LEGACY FALLBACK: PRESERVE EXISTING PROVEN ACTION WHEN UNCERTAIN SCENE STYLE: passionate, explicit, physically natural and continuity-focused INTENSITY: PASSIONATE ACTION COMMITMENT: HIGH RHYTHM: STRONG CONTROLLED ESCALATION SOFT TISSUE PHYSICS: REALISTIC INERTIA: BODY-DRIVEN DAMPING: NATURAL RESPONSE SCALE: MOTION-MATCHED OSCILLATION LIMIT: SINGLE DAMPED SETTLE CONTACT DEFORMATION: ON CONTACT ANCHOR: SURFACE-LOCKED VOLUME PRESERVATION: ON GRAVITY RESPONSE: ON ENDING: CONTROLLED BY THE ENDING MODE SELECTOR CAMERA: PRESERVE VISIBLE NUDITY AUTO-TRIGGER: ON DIALOGUE: none unless clearly requested AUDIO: natural breathing, vocal reactions and environmental sound only; no music PERFORMANCE STYLE: LIFELIKE AND REACTIVE BODY LANGUAGE: WEIGHT-AWARE, ASYMMETRIC, CAUSAL FACIAL PERFORMANCE: CONTEXTUAL MICRO-REACTIONS MOTION TIMING: ORGANIC, NON-METRONOMIC AUDIO SYNC: ACTION-BOUND DIEGETIC SUPPORTING MOTION BUDGET: 1-2 RELEVANT CUES PER BEAT POSITION RECOGNITION: AUTO-DETECT FROM VISIBLE GEOMETRY POSITION MODE: PRESERVE CURRENT WHEN ACTIVE POSITION TRANSITION: MAXIMUM ONE LOW-COST TRANSITION ROLE ASSIGNMENT: GEOMETRY-BASED, ANATOMY-OWNER LOCKED MIXED ANATOMY RECOGNITION: STRICT LITERAL VISUAL EVIDENCE FUTANARI ANATOMY LOCK: PRESERVE BREASTS + ATTACHED PENIS PER OWNER PERSISTENT ANATOMY EXISTENCE: PRESENT_LOCKED SURVIVES OCCLUSION + RECEIVING ROLE OCCLUDED HAND COMPLETION: FORBIDDEN — PRESERVE VISIBLE SILHOUETTE ACTION CLASSIFICATION: SEPARATE EFFECTOR + TARGET FROM POSITION ACTION LOCK: PRESERVE MANUAL / ORAL / PENETRATIVE / OTHER CLASS POSITION VARIANT: AUTO-DETECT INDEPENDENTLY CONCURRENT ACTIONS: AUTO — ONE PRIMARY + ONE COMPATIBLE SECONDARY GENITAL REINTERPRETATION: FORBIDDEN The user has activated the separate graph-level confirmation that every depicted participant is 18 or older and consenting. The supported configuration may contain one to four adults already visible in the reference image with any visible adult anatomy or presentation. Treat the image as the literal first frame. Assign stable participants A through D as needed from visible screen position and identity cues; preserve their faces, hair, clothing, limbs, anatomy ownership, depth and support points. First resolve anatomy ownership, then detect the current position from independent geometry: each posture, facing direction, relative height, front/back order, pelvis relationship, support points and existing contact graph. Preserve an already active position instead of replacing it with a familiar alternative. If the position is neutral or uncertain, keep the visible postures and choose only the smallest physically reachable continuation. Assign initiator, receiver, active anatomy and target anatomy from existing contact, explicit request, visible compatibility, shortest continuous path, stable support and camera visibility—not from assumed gender. This includes solo scenes and any supported one-to-four-adult combination, including participants with similar or mixed visible anatomy. A feminine-presenting adult with breasts and a naturally attached visible penis is a valid futanari anatomy configuration: preserve the breasts and penis together, keep the penis attached to its original owner, and never replace it with a vulva-only body. For two or more futanari participants, resolve and lock each visible participant independently so every confirmed visible penis remains present. Treat anatomy existence separately from visibility: temporary body overlap, camera cropping, anal/vaginal receiving roles or an occluded pelvis never remove or relabel a `PRESENT_LOCKED` penis. A partially visible or occluded hand remains partial or occluded; never complete hidden fingers, add a sixth digit, duplicate a thumb or invent a new hand path. Never exchange, duplicate, erase, normalize or reinterpret anatomy. Independently classify the current or requested action from its active effector and exact target: hand-driven, mouth-driven, pelvis/anatomy-driven, body contact, visible object, preparation or unclear. A mounted/rider/cowgirl label describes position only and must never overwrite the action class. Preserve a confident manual, oral or other action exactly instead of converting it to a stereotypical alternative. Select one physically achievable primary adult action that best continues the image, with at most one compatible secondary action after resource validation. Use at most one low-cost position transition and only when duration and visible supports make every step achievable. Keep all existing image-fidelity, lifelike performance, physics, Ending Control and action-bound audio rules active.	—
adult_confirmed	BOOLEAN	false	—
ollama_model	STRING	fredrezones55/Qwen3.5-Uncensored-HauhauCS-Aggressive:9b	—
ollama_url	STRING	http://127.0.0.1:11434	—
ollama_context_profile	COMBO	8-12 GB	3 options: 8-12 GB, 16 GB, 24-32+ GB
ending_mode	COMBO	AUTO	4 options: AUTO, NO CLIMAX, CLIMAX, LOOP / CONTINUOUS ACTION
imageopt	IMAGE		—
ltx_prompt_profileopt	COMBO	LTX 2.3	2 options: LTX 2.3, LTX 2.5

Outputs (3)

Name	Type	Description
prompt_package	VELVET_VICE_PROMPT_PACKAGE	—
final_prompt_preview	STRING	—
status	STRING	—