Show HN: Kling O1–Unified Multimodal Video Model for Consistent Video Generation

Hi HN,

We’ve been building Kling O1, a unified multimodal video model that handles nearly every major video-generation task—reference-to-video, text-to-video, start/end frames, editing, transformations, restyling, and camera extension—all inside one system.

Key Features of Kling O1

• Input Anything – One Model for All Tasks

Images, clips, characters, layouts, and text instructions all flow into a single Kling O1 pipeline. No switching tools.

• Multimodal Understanding

Everything you upload becomes part of the “prompt.” Kling O1 interprets subjects from multiple viewpoints to generate accurate motion and fill missing frames with precision.

• All-in-One Reference for Consistency

Feed Kling O1 reference images or video, and it preserves characters, props, and scenes across shots—solving the long-standing video-consistency problem.

• Stack Multiple Edits at Once

Add subjects, change backgrounds, restyle, and apply element-based controls in a single generation.

• 3–10 Second Shots for Storytelling

Control pacing with flexible clip lengths. Start/End Frame control is coming soon.

Where Kling O1 Is Useful

Advertising – Upload product, model, and background images + a short prompt to generate polished product shots quickly. Fashion – Build endless virtual runway videos from models + clothing refs.

Film Post-production – “Remove the bystanders” or “make the sky blue”—Kling O1 handles pixel-level corrections via language.

Filmmaking – Lock in characters, props, and scenes for consistent multi-shot sequences.

We’d love feedback on:

• Which parts of the unified workflow help you most

• What multimodal controls you’d want next

• Scenarios where 3–10 second shots or start/end frames would be most useful

Thanks for reading!

Try Kling O1 here: http://klingo1ai.net/?i=d1d5k