Omniscribe: fully local video/GIF/image/audio captioner on 8GB VRAM, MIT licensed
official_imvoiid · reddit · 2026-10-18
A developer released Omniscribe, a portable fully local captioner for video, GIF, image and audio (it hears sound too), with batch mode for training sets. Runs on 8GB VRAM on Windows under MIT license, plus a MiniMax H3 prompt maker.
More from Multimodal
- Ask Gemini to make a meme about AI and it nails the Anakin-Padmé template — SoleilMartin8595 · 2026-10-18
- Fan-made AI short film reimagines Death Note with cinematic results — neptune765 · 2026-10-18
- krea2 LoRA training: 40-60 min on RTX 5090 vs 12 min on Fal.ai — Ambitious-Equal-7141 · 2026-10-18
- Why MOSS Image looks better: it deliberately excluded AI images from training data — ostrisai · 2026-10-18
- 10-year visual artist shares hybrid AI workflow showreel: ComfyUI, Midjourney, Blender — neptune765 · 2026-10-18
- The fill-in-the-blank prompt: Claude as motion director and Remotion coder, end to end — aziz4ai · 2026-10-18