← Workflows

InfiniteTalk Image-to-Video (lip-sync)

Image → Video

Audio-driven talking-head. Animates a still image to lip-sync a voice line (Wan 2.1 I2V backbone + InfiniteTalk/MultiTalk). Talking-head motion only, not big body motion.

infinitetalk_i2v12 GB VRAM19.8 GB of weights

Calling it

FlixML has no generate button: you tell your agent what you want and it calls the API on your install. The id is what names this workflow.

POST /api/video/generate
{
  "workflow": "infinitetalk_i2v",
  "prompt": "...",
  "image": "...",
  "audio": "..."
}

Parameters

promptstrrequired

Drives expression/motion. InfiniteTalk is talking-head; keep it simple.

imagestrrequired

Input image filename (staged to ComfyUI input)

audiostrrequired

Input audio filename (wav) staged to ComfyUI input. Drives lip-sync + clip length.

negative_promptstrdefault bright tones, overexposed, static, blurred details, subtitles, style, works, paintings, images, static, overall gray, worst quality, low quality, JPEG compression residue, ugly, incomplete, extra fingers, poorly drawn hands, poorly drawn faces, deformed, disfigured, misshapen limbs, fused fingers, still picture, messy background, three legs, many people in the background, walking backwards
widthintdefault 640

Resize target width (divisible by 16)

heightintdefault 640

Resize target height (divisible by 16)

lengthintdefault 81

Frame count. At 25fps: 81 frames ~= 3.2s. Keep short — long renders choke 12GB/16GB RAM.

fpsintdefault 25

InfiniteTalk is trained at 25fps; do not change unless you know why.

seedintdefault 2
stepsintdefault 6

Sampler steps. 6 with the lightx2v speed LoRA is the tuned default.

cfgfloatdefault 1
shiftfloatdefault 11
schedulerstrdefault dpm++_sde
blocks_to_swapintdefault 12

Block-swap count for low-VRAM. 12 is the proven sweet spot on a 12GB card with the Q3_K_S GGUF: ~10GB VRAM used, GPU at 95%, no OOM. Higher = less VRAM but GPU-starved/slow; lower risks OOM.

modelstrdefault wan2.1-i2v-14b-480p-Q3_K_S.gguf

GGUF-quantized Wan2.1 I2V 14B backbone (~7.9GB). Fits 12GB card with block-swap. fp8 safetensors OOMs a 12GB card.

multitalk_modelstrdefault Wan2_1-InfiniteTalk_Single_Q4_K_M.gguf

GGUF-quantized InfiniteTalk module (~1.3GB). Must be GGUF to pair with the GGUF backbone.

wav2vec_modelstrdefault wav2vec2-chinese-base_fp16.safetensors
t5strdefault umt5-xxl-enc-fp8_e4m3fn.safetensors
vaestrdefault Wan2.1_VAE.safetensors
clip_visionstrdefault CLIP-ViT-H-14-laion2B-s32B-b79K.safetensors
speed_lorastrdefault WanVideo\Lightx2v\lightx2v_I2V_14B_480p_cfg_step_distill_rank64_bf16.safetensors
speed_lora_strengthfloatdefault 1
filename_prefixstrdefault infinitetalk_i2v

Your install is the source of truth: GET /api/workflows returns these parameters as they are on your machine.

Same input, different job

Wan 2.2 Image-to-Video

Image-to-video with Wan 2.2 — animate a still into a short clip, with a motion prompt driving the movement. Identity and scene come from the source image.

Wan 2.2 Image-to-Video (context windows)

Image-to-video past the model's 81-frame limit, generated as ONE clip instead of a chain. ComfyUI samples the whole length in overlapping context windows and fuses them, so the action carries across the joins — use this instead of taking a clip's last frame and animating it again, which hands the model a still and restarts the motion from rest.