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
feat(panel): Characters tab — extra LoRAs, i2v, i2v+audio, 15s, progress bar fix
aafbb53 mrbizarro committed 6 months ago
refactor(panel): Characters tab is now a thin wrapper around /queue/add
7763034 mrbizarro committed 6 months ago
feat(panel): Characters strip-down + Draft mode + Codex recipe replay + captions UI + Style mode
b3f3652 mrbizarro committed 6 months ago
fix: gallery exclusion bug + aggressive_cleanup on pipeline rebuild (#7)
934260f colinbdesign committed 6 months ago
feat(panel): auto-use first training image as character avatar
d6a11fb mrbizarro committed 6 months ago
fix(panel): voice training bugs surfaced by first end-to-end run
126b844 mrbizarro committed 6 months ago
feat(panel): Train-tab dataset guidance + Turbo toggle in Characters
3008270 mrbizarro committed 6 months ago
feat(characters): Load Params routes back to Characters when applicable
2dd7a1c mrbizarro committed 6 months ago
feat(panel): tighter two-column layout — composer left, cinematic stage right
c3c24d6 mrbizarro committed 6 months ago
feat(helper): plumb video_skip_step + audio_skip_step into HQ pipe
da87af1 mrbizarro committed 6 months ago
feat(characters): new Characters tab — pick character, prompt, ship
3ad6401 mrbizarro committed 6 months ago
feat(panel): finish agent UI rip-out + larger OKLCh wordmark gradient
702986e mrbizarro committed 6 months ago
fix(helper): fps→frame_rate kwarg shims + drop _phosphene_dit_fused latch
aa235aa mrbizarro committed 6 months ago
feat(image-studio): remove Flux/Z-Image, enable FBCache on all Qwen tiers
9b1f8b2 mrbizarro committed 6 months ago
feat(qwen): FBCache for Qwen-Image-Edit via mflux patch — 1.39x faster
6faa500 mrbizarro committed 6 months ago
fix(install): demote one-stage fps patch to advisory on upstream drift
a2fe426 mrbizarro committed 6 months ago
feat(image-studio): standardize Qwen + Flux tiers — Fast / Medium / Quality
35205a3 mrbizarro committed 6 months ago
fix(image-studio): real time estimates — was showing ~30 min for Fast
824c36f mrbizarro committed 6 months ago
fix(update.js): use uv pip install for smolagents/hub step — was failing every update
4d4341b mrbizarro committed 6 months ago
fix(image-engine): strip Malloc* env vars on subprocess spawn
0d7cb13 mrbizarro committed 6 months ago
feat(image-studio): ✦ Quality button on photo cards — remake at Quality
25959c3 mrbizarro committed 6 months ago
feat(image-studio/hidream): 3 modes — Fast / Medium / Quality
440ba02 mrbizarro committed 6 months ago
perf(image-studio/hidream): 6-step + FBCache 0.15 — quality + ~80s
10c3eda mrbizarro committed 6 months ago
perf(image-studio/hidream): 3-step flow_match — 5min -> ~1min HD edits
5d9c8c9 mrbizarro committed 6 months ago
perf(image-studio/hidream): drop default steps 28 -> 20 (~5 min HD edits)
ed04005 mrbizarro committed 6 months ago