Newtake Launches in Korea With East Asian Character Consistency AI Video Workflow

On September 11, 2026, LiblibAI (operating under the global brand LibTV) officially launched Newtake, an AI video production platform built on an infinite canvas and modular node-based workflow, selecting South Korea as its primary strategic launchpad. Aimed at addressing persistent challenges in generative video production, the platform focuses on natural East Asian character rendering and robust character consistency pipelines tailored for multi-shot commercial productions.

Image source: X @ProCreatorHarry / Newtake

Unlike conventional web-based tools centered on one-off text-to-video prompt generations, Newtake unifies the entire production lifecycle into a single collaborative workspace. The platform covers pre-production character asset definition, multi-angle camera perspective parsing, and third-party foundation model rendering, targeting professional workflows across web dramas, long-form episodic series, and commercial brand advertising.

Infinite Canvas and Node-Based Modular Production Environment

At the core of Newtake is an infinite canvas paired with a node-based visual workflow designed for high-density production.

Creators can arrange distinct production phases—such as text prompting, reference image curation, motion control parameters, character sheets, and upscale rendering—as interconnected visual nodes across an expansive workspace. This setup eliminates the need to switch between disconnected web applications, download intermediary video clips, and re-upload files to disparate tools.

• Non-Linear Visual Workspace: Filmmakers can inspect their entire sequence from storyboard layout to final render outputs within a single overview, branching scenes flexibly.
• Reusable Node Architecture: Core functional components, including prompt recipes, style anchors, character sheets, and rendering engines, can be packaged into reusable node graphs.
• Centralized Asset Management: Visual assets and render versions remain organized within their contextual nodes, preserving full traceability throughout complex production pipelines.

According to the developers, South Korea was chosen as the initial target market due to its dynamic creator ecosystem and high demand for episodic short-form drama and webtoon video adaptations, providing critical real-world feedback to refine the platform.

Textured Character Sheets and Multi-View Parsing for Identity Consistency

In professional generative video production, identity drift—where a character's facial features, wardrobe, or aesthetic proportions shift unpredictably between cuts—remains a major hurdle. Newtake introduces specialized structural mechanisms to enforce visual consistency across multi-scene projects.

The platform specifically emphasizes enhanced rendering of East Asian facial contours, gaze alignment, and authentic skin textures. By addressing the rigid posture and artificial synthetic skin tones often produced by Western-centric base foundation models, Newtake aims to generate more naturalistic human expressions.

• Textured Character Sheets: Creators lock down front, side, and three-quarter views of a character alongside wardrobe specifications onto high-resolution reference sheets that serve as immutable identity anchors.
• Multi-View Parsing: The system programmatically derives corresponding camera angles—including frontal, lateral, profile, high-angle, and low-angle frames—from unified character and background source assets.
• Cross-Cut Visual Continuity: Characters retain consistent facial geometry, skin realism, and stylistic harmony even when subjected to dynamic camera movements, varied lighting setups, and distinct background scenes.

This structured approach significantly cuts down on trial-and-error prompting across individual shots, allowing creators to direct scenes with the precision required for narrative film and commercial campaigns.

Third-Party Foundation Model Integrations and Practical Production Caveats

Rather than relying strictly on an isolated proprietary generation engine, Newtake integrates leading third-party video foundation models directly into its node graph environment.

The current lineup includes ByteDance's Seedance 2.5, MiniMax H3, Kling 3.0, and Wan 3.0. Directors can select and route different models on a per-node basis, evaluating each model's distinct physical dynamics simulation, motion controllability, and cinematic texturing to find the optimal aesthetic match for specific narrative beats.

However, creative teams exploring commercial deployment must account for practical operational constraints and cost structures.

• Third-Party Model Credits and Paid Subscriptions: Calling advanced models such as Seedance 2.5 or MiniMax H3, as well as initiating high-resolution upscaling, consumes additional platform credits beyond basic generation quotas.
• Node Interface Learning Curve: While offering superior architectural control, the infinite canvas paradigm demands foundational training in node parameter routing compared to simple single-prompt web generators.
• Commercial Licensing and Asset Rights: Production houses must thoroughly audit platform usage terms regarding commercial rights and credit consumption across complex, multi-shot sequences before full deployment.

By unifying East Asian character fidelity, structural identity controls, and high-performance foundation model routing, Newtake presents a compelling next step toward cohesive, commercial-ready AI video production pipelines.

Sources

• Newtake Official Website: Newtake AI Video Creation Platform — Official portal and production workspace for LiblibAI/LibTV's infinite canvas and modular node-based AI video creation suite.
• @ProCreatorHarry Announcement on X: Newtake Launch in Korea and Character Consistency Workflow — Practical breakdown and demonstration video showcasing East Asian character consistency, identity preservation, and multi-shot continuity in a real-world Korean production environment.