Fish Audio Drama 3 clones voices, casts 16-character scenes and takes direction. Try it free in a Hugging Face Space, plus the license and 48 GB GPU catch.
Iris-3B Turbo is a free, Apache 2.0 4-step AI image generator with no VAE, plus depth and 4x upscaling. Try it in a free demo or run it locally.
Qwen-Image-2.1-Turbo edits and generates 2K images in 8 steps. Try it free in a browser or on a 12 GB GPU, and know the non-commercial license catch.
Prism is a free, MIT-licensed image-to-video model that adds speech, sound effects and music at up to 2K. Here's how to try it free without an 80 GB GPU.
HyperFrames v0.8.145 adds GPU export. HeyGen's free Apache-2.0 tool lets your AI agent turn HTML into MP4 motion graphics with no render fees. Setup inside.
Suno Speech makes narration and music in one take, free to try. Here's how to use it, and how to build an open-source version with Kokoro, ACE-Step and ffmpeg.
Google's Nano Banana 2.1 halves API image prices but has no free API tier. Here's how to try it free, plus FLUX.2 klein 4B, the Apache 2.0 alternative.
Sopro V2 Turbo is a free, Apache-2.0 voice cloner that copies a voice from 5-20 seconds of audio and runs on a laptop CPU or right in your browser.
DMAD makes MiniMax-H3 generate video with stereo sound in 4 steps. Try the free demo, run it on a 24 GB GPU, or use the license-safe alternatives.
Inkvec is a free, Apache-2.0 AI logo vectorizer that turns PNG and JPEG logos into SVG in your browser, plus Agate, a 261M image model you can run in a tab.