Free, open-source AI video generator for Mac. Run LTX-2.5 and MiniMax H3 (Hailuo) locally on Apple Silicon: text-to-video and image-to-video with sound, lip-sync, character LoRA training, songs from lyrics (YuE2) and image editing (Qwen-Image-Edit). One-click install via Pinokio. Offline: no cloud, no API key, no NVIDIA GPU.
Python ai-filmmakingai-videoai-video-generatorapple-siliconhailuoimage-to-videolip-synclocal-ailora-trainingltx-2ltx-2-5ltx-videomacosminimax-h3mlxmusic-generationoffline-aipinokiotext-to-videovideo-generation
Commits on main
release(v3.4.0): Hailuo H3 — a second video engine, with voice
0ef1a98 mrbizarro committed 3 months ago
release(v3.3.0): Remix wave — Ingredients, Control, Colorize, Ingredients×Character, A2V + a sample character
733e37f mrbizarro committed 3 months ago
fix(install): fail fast on macOS <14 + name the Metal GPU-timeout abort (#44)
fbfa56b mrbizarro committed 4 months ago
fix(image): Qwen-Image-Edit install hint pointed at a renamed menu + a fake button
5cfea3e mrbizarro committed 4 months ago
fix(stats): quiet + cosmetic dashboard-fetch failures (#43)
485ff49 mrbizarro committed 4 months ago
fix(train): precision-aware dev-transformer selection — quantized dev can't shadow the full one (#35/#36)
5be9e47 mrbizarro committed 4 months ago
release(v3.2.6): pin transformers <5.13.0 — fixes the crash on every generation
0f6b21b mrbizarro committed 4 months ago
release(v3.2.5): image-to-image + Ideogram reference bridge, robust first-run downloads
fbc6335 mrbizarro committed 5 months ago
release(v3.2.4): Ideogram runs on slower Apple GPUs (M1/M2)
edd84dc mrbizarro committed 5 months ago
release(v3.2.1): Ideogram 4 out of the box — un-gated mirror + mflux bundled
c46d879 mrbizarro committed 5 months ago
release(v3.2.0): Ideogram Layout editor — on-canvas design tool + agent API
6117225 mrbizarro committed 5 months ago
release(v3.1.0): Ideogram 4 — text-rendering image model + visual text-placement
e25e433 mrbizarro committed 5 months ago
release(v3.0.13): character LoRA fixes — spaced-filename classification + Character-tab desync
814dc2d mrbizarro committed 5 months ago
fix(character): spaced-filename classification + Character-tab desync (#5)
ce03139 mrbizarro committed 5 months ago
release(v3.0.12): real mosaic fix (deep checksum verify) + #5 LoRA + cgi→email + multi-keyframe UI
80752a7 mrbizarro committed 5 months ago
fix(helper): drain stranded ready event from _read_until + runtime fingerprint (v3.0.11)
093a9f0 mrbizarro committed 5 months ago
feat(integrity): detect corrupt/partial model weights + one-click Repair (mosaic self-heal)
173d2fd mrbizarro committed 5 months ago
release: v3.0.9 — custom I2V size restored + Train dev-transformer Download fix
ac0461a mrbizarro committed 5 months ago
fix(train): dev-transformer Download → 'unknown install key' (multipart vs parse_qs)
7781c03 mrbizarro committed 5 months ago
fix(panel): restore custom Width×height inputs for Image-to-Video
0b98a5b mrbizarro committed 5 months ago
release: v3.0.8 — promote ltx-2-mlx v0.14.8 catch-up to public
0dc63d2 mrbizarro committed 5 months ago
docs(STATE): record the v0.14.8 ltx-2-mlx catch-up + pin standing rule
add7cd3 mrbizarro committed 5 months ago
ltx-2-mlx: catch-up pin v0.14.0 -> v0.14.8, drop 6 of 7 runtime patches
aaf1ba8 mrbizarro committed 5 months ago
fix(docs/infra): correct ltx-2-mlx pin in CLAUDE, scenema URL quoting, ignore backups, VERSION 3.0.7 (v3.0.7 P2/P3)
74d2df4 mrbizarro committed 5 months ago
fix(panel): close GPU-contention inverse race + surface ltx version in /status (v3.0.7 P1/P2)
b069e2d mrbizarro committed 5 months ago
fix(comments): un-mislead two load-bearing code comments (doc audit H7/H8)
c139a2f mrbizarro committed 5 months ago