Home music studio with monitor speakers and mixing desk, where creators can run YuE2 locally for free
YuE2 now runs free on an 8 GB GPU, an Apple Silicon Mac or in ComfyUI. File sizes, VRAM, exact commands, and the non-commercial license catch.
Hand-drawn website wireframe sketch on a notebook next to a smartphone, representing AI design editing with Ideogram 4.5
Ideogram 4.5 edits text and details without drift. Here is what the free plan gets you, real arena scores, and the free Apache-2.0 editor to run locally.
Home recording studio microphone and headphones for AI voiceover work
Eleven v4 tops the independent TTS arena. Here's how to try ElevenLabs' new AI voice model free today, plus Apache and MIT open-source alternatives.
Cinema camera on a film set, illustrating free AI video restoration with LTX-2.5 Restore and Refine Details
Lightricks' free LTX-2.5 Restore and Refine Details LoRAs clean, colorize and sharpen video. Specs, the 32 GB VRAM reality, ComfyUI setup and a free demo.
Mirrorless camera on a white background, illustrating AI camera-angle control with Krea 2 Anygles
Krea 2 Anygles is a free LoRA that re-renders a person from any camera angle using one image. Try the no-GPU demo, ComfyUI setup, VRAM needs and license limits.
Bright modern living room with sofa, coffee table and staircase, the kind of single photo Mira-Scene turns into a 3D scene
Mira-Scene turns one photo into an editable image to 3D scene with separate objects. Here is what it needs, the license catch, and free browser demos to try today.
Camera monitor showing a performer on set, illustrating reference footage used to guide AI video ControlNet generation
MiniMax-H3 ControlNet Union 2.0 adds 8 AI video controls, but its license excludes the US. Here's the free Apache-2.0 Wan 2.2 Fun Control path for 8 GB GPUs.
Podcaster recording voiceover at a microphone, the kind of work Gemini 3.8 Flash TTS voice design targets
Gemini 3.8 Flash TTS designs new AI voices from a text prompt and is free in Google AI Studio. Pricing, limits, code, and the open Qwen3-TTS alternative.
Hand-drawn website wireframe sketch in a notebook next to a smartphone, representing AI UI design generation
Ming-Image-0.1-Design is a free, MIT-licensed AI model for UI mockups, posters and transparent PNGs, now native in ComfyUI. Demo, files, VRAM and trade-offs.
Professional music recording studio with mixing console, studio monitors and acoustic panels, representing free AI music generation with DiffSynth-Music
DiffSynth-Music adds beat, vocal, accompaniment and melody control to ACE-Step 1.5. It's Apache 2.0 with a free demo. Here's the setup, specs and catches.