Traditional 2D Animation LoRA for LTX-2.3, trained on curated public-domain clips to push hand-drawn Western cel-animation linework, readable silhouettes, elastic motion, subtle contour wobble, and classic frame-by-frame timing. Best for stylized cartoon action, expressive acting, and vintage-inspired 2D motion language rather than strict franchise replication. You can use this prompt generator which has been customized to work alongside this LoRA: https://chatgpt.com/g/g-69dd480de1308191ae3e7e959e8ecfc0-ltx-2-3-traditional-2d-cel-animation-prompter Use with LTX-2.3: This LoRA is built to reinforce traditional Western 2D cel-animation linework, readable silhouettes, elastic motion, subtle contour wobble, and classic frame-by-frame timing inside LTX-2.3. Best results: Use prompts that focus on clear action, strong silhouettes, simple readable staging, and classic animation principles like anticipation, squash and stretch, overlap, follow-through, recoil, and expressive facial acting. Core activator: "cellinework, traditional western 2d animation, inked outlines, stylized cartoon linework, medium line weight, subtle contour wobble, elastic drawn motion, subtle handmade flicker, continuous shot, no cuts" Recommended prompt flow: Subject and shot setup, simple setting, cel-animation activator, chronological action beats, camera direction, continuity constraints. Works best for: T2V action shots, I2V animation from strong stills, walk cycles, expressive acting, prop interaction, and smear-frame or snap-accent motion tests. Prompt modes: finished cel for polished inked outlines and cel-painted finish. rough pencil-test for sketchy animator-drawing motion with broken contours and graphite feel. black-and-white vintage for monochrome theatrical cartoon treatment. Camera guidance: Best with static shots, simple pans, gentle push-ins, pull-backs, side tracking, and clean tilts. Avoid shaky handheld, chaotic reframing, or modern live-action camera language. I2V guidance: Start from the exact reference image, preserve design, proportions, palette, and staging, then add only small readable motion beats. Refine / low-denoise passes: Keep the exact same shot, timing, characters, props, and action. Use later passes only to reinforce line clarity, silhouette stability, facial consistency, and cel-fill readability. Avoid: Photorealism, live action, CGI, 3D render, anime or manga wording, muddy outlines, random camera shake, scene changes, and overly franchise-specific style language unless intentionally testing recognition. General note: This LoRA works best as a strong style influence rather than a total override, so simpler prompts with clearer motion usually perform better than overly dense cinematic descriptions. ___________________________________________________________________________________ Here are two prompt examples to be used with the reference images provided: I got sidetracked by the wrong tool. Here are the prompts. 2D Lora Test Reference 01.jpg Positive Prompt: I2V, 81f, 16:9, start from the exact reference image, preserve the same elderly man’s design, same balding gray hair, heavy brows, long pink nose, green turtleneck sweater, same head proportions, same face shape, same blue background, same close-up composition. cellinework, traditional western 2d animation, inked outlines, stylized cartoon linework, medium line weight, subtle contour wobble, elastic drawn motion, subtle handmade flicker, continuous shot, no cuts. The man begins in his current speaking pose, then gives a sly side-eye glance, lifts one eyebrow, narrows the eyes, and opens his mouth through clear dialogue shapes as if delivering a dry clever remark. Midway, he tilts his head slightly forward, the brows rise and settle, the mouth stretches into a smug half-smile, and the cheeks and jaw shift with readable facial acting. Toward the end, he closes his mouth, blinks once, and settles into an amused knowing expression. Static camera, no new props, no background change. Negative Prompt: photorealism, live action, cgi, 3d render, anime, manga, off-model face, identity drift, extra features, broken anatomy, duplicated eyes, fused features, muddy outlines, blurry linework, background drift, camera movement, zoom, pan, cut, scene change, text, watermark, logo, stiff motion, weak facial acting, loss of cel animation style 2D Lora Test Reference 02.jpg Positive Prompt: I2V, 81f, 16:9, start from the exact reference image, preserve the same bear and rabbit designs, same forest path setting, same foliage, same action staging, same proportions, same colors, same angle, same chase composition. cellinework, traditional western 2d animation, inked outlines, stylized cartoon linework, medium line weight, subtle contour wobble, elastic drawn motion, subtle handmade flicker, continuous shot, no cuts, strong anticipation, follow-through, drag, squash and stretch. The rabbit keeps sprinting forward in panic with rapid foot taps, flaring ears, and a bouncing red hat that lags behind each stride. The bear lunges after the rabbit with heavy bounding steps, claws reaching forward, mouth open in a furious shout, shoulders and belly showing strong weight and recoil. Midway, emphasize hand-drawn smear frames on the fastest leg and arm actions, readable chase arcs, dirt kick-up, tail and ear overlap, and elastic impact timing. Toward the end, the rabbit snaps into a sharper burst of speed while the bear overreaches, rebounds, and continues the chase. Static camera, preserve the original chase direction and forest layout. Negative Prompt: photorealism, live action, cgi, 3d render, anime, manga, off-model bear, off-model rabbit, extra characters, duplicated limbs, broken anatomy, muddy outlines, blurry linework, random camera shake, pan, zoom, cut, scene change, background drift, prop drift, heavy realistic motion blur, digital tweening, vector-clean motion, text, watermark, logo, loss of cel animation style If you want, I can turn these into versions for finished cel, rough pencil-test, or black-and-white vintage mode.

Style Lora