Phosphene 2.0 Comfortable 13.0 / 48 GB · swap 0.3 helper warm models 3/3 queue 0 idle ✓ Up to date · 2.0.5 by Bizarro ↗ X Mode Text prompt → video Image image + prompt FFLF first + last frame Extend continue a clip Quality Quick 640×480 · ~2 min Q4 · any Mac Balanced 1024×576 → 720p · ~5 min Q4 · ~5 min · no crop Standard 1280×704 · ~8 min Q4 · standard tier+ High 1280×704 · ~12 min Q8 + TeaCache Prompt Describe the scene AND the sound: e.g. 'wizard in a forest clearing, fireflies spiraling up — low whispered chant, ember crackle, distant owl'. Audio is generated jointly with video; without sound cues the model outputs near-silent ambient. Avoid Optional: blurry hands, distorted fingers, extra fingers, smeared face, warped text ✨ Enhance with Gemma Drop .safetensors files into /Volumes/MicheleExt4TB/pinokio/api/phosphene.git/mlx_models/loras to use them, or click Browse CivitAI above. Each LoRA picks up an optional sidecar .json with name + trigger words + recommended strength. Advanced Duration (s) 5 Frames (8k+1) 121 Seed (-1 random) -1 Duration 5.00s @ 24fps · 1024×576 → 1280×720 fit · Steps 8 Generate Stop Batch paste Pause queue Clear queue No outputs yet generate something on the left to begin Outputs · 0 Visible Hidden Now Queue Recent Logs ltx-2.3-mlx-q4 Collapse [13:53:16] caffeinate active — Mac won't idle-sleep while queue is running [13:53:16] Run via helper: id=j-19e30a230ff-001 mode=t2v quality=quick accel=off 640x480 121f [13:53:16] Spawning warm helper (low_memory=true, idle_timeout=1800s) [13:53:16] helper ready · model=/Volumes/MicheleExt4TB/pinokio/api/phosphene.git/mlx_models/ltx-2.3-mlx-q4 · low_memory=True [13:53:16] step:get_pipe kind=t2v [13:53:16] Loading T2V pipeline (first job is the slow one)... [13:53:16] step:get_pipe done [13:53:16] accel:mode off [13:53:16] step:generate mode=t2v 640x480 121f @24.0fps steps=8 accel=off [13:53:16] [Loading text encoder (Gemma)] ... [13:53:20] [Loading text encoder (Gemma)] done in 3.8s [13:53:25] [Loading transformer (transformer-distilled.safetensors)] ... [13:53:25] [Loading transformer (transformer-distilled.safetensors)] done in 0.2s [13:53:42] Denoising: 0%| | 0/8 [00:00

@mfadda