Documented core-node templates with explicit setup requirements. Download or open in Studio.
Loading local templates…
Checkpoint, positive and negative conditioning, empty latent, sampling, VAE decoding and save.
Requires a compatible installed checkpoint. Intended for conventional checkpoint models; this is not a Flux diffusion-loader graph.
Verification: exported references checked by OnBeg. Runtime: not executed on your installation.
Checkpoint, input image, VAE encoding, conditioning, sampling, decode and save.
Requires a compatible installed checkpoint and an image in your ComfyUI input directory. Intended for conventional checkpoint models; this is not a Flux diffusion-loader graph.
Verification: exported references checked by OnBeg. Runtime: not executed on your installation.
Separate diffusion loader, matching CLIP-L and T5 encoders, Flux guidance and model shift, 16-channel latent, sampler, VAE and image output.
Choose your installed diffusion model, encoders and matching VAE in Studio. GGUF files require ComfyUI-GGUF. Model weights are not included or downloaded automatically.
Verification: graph references checked. Runtime on your installation requires a real test.
Separate diffusion loader, matching CLIP-L, CLIP-G and T5 encoders, SD3 model shift, 16-channel latent, sampler, VAE and image output.
Choose your installed diffusion model, encoders and matching VAE in Studio. GGUF files require ComfyUI-GGUF. Model weights are not included or downloaded automatically.
Verification: graph references checked. Runtime on your installation requires a real test.
For model-specific examples including Flux and video families, use official ComfyUI workflow templates. We do not label unexecuted community graphs “error-free”.