Nodes/ComfyUI_mittimiLoadPreset/LoadAndSettingParameters01
ComfyUI Node

LoadAndSettingParameters01

The integrated preset node that lets tweaks stick

By mittimi·Created 2 years ago·Updated 2 years ago· 1
LoadAndSettingParameters01
    • POS A
    • POS C
    • NEG A
    • NEG C
    • ckpt_name
    • stop_at_clip_layer
    • vae
    • Steps
    • CFG
    • sampler_name
    • scheduler
    preset
    PosPromptA
    PosPromptC
    NegPromptA
    NegPromptC
    checkpoint_name
    ClipNum-1
    vae_name
    Steps
    CFG
    SamplerName
    Scheduler

    This is the version that answers the complaint the split design creates: "I fixed the clip skip and it reverted anyway." LoadAndSettingParameters01 merges the preset picker and the parameter display into a single node, and it changes when the preset applies - which is the whole point.

    What's different from the two-node setup

    In the split design (LoadPresetForSetting01 + SettingParameters01), the preset is re-read at queue execution, so manual edits get overwritten. Here, the preset only rewrites the widgets at the moment you pick it from the dropdown. After that, whatever's on the widgets is what runs. Tweak ClipNum from -2 to -1, bump CFG a bit, fine-tune the prompt - it sticks. The README positions this as the fine-tuning-friendly option, and it earns that.

    The tradeoff the author is honest about: the node auto-selects a preset when it's created. That means if you drag an image back into ComfyUI to rebuild the workflow from its embedded metadata, the node re-applies the preset and wipes your values. That's the exact scenario where the split version behaves better. So the rule of thumb: live-tweaking one graph in front of you → use this; repeatedly loading saved workflows → the split version's predictable queue-time behavior is what you actually want.

    How it works

    Same websocket handshake as the rest of the pack: the Python side pushes a my.custom.message to the frontend, js/web.js rewrites the widget values, and the node returns whatever the widgets show. The preset fires when the dropdown value changes, not on every queue.

    The output roster matches SettingParameters01:

    • POS A / POS C / NEG A / NEG C - four text prompts for your CLIPTextEncode nodes.
    • ckpt_nameCheckpointLoaderSimple's ckpt_name input.
    • stop_at_clip_layer (INT) → CLIPSetLastLayer. Negative numbering, default -1; -2 is the classic anime/Pony value.
    • vae - a genuinely loaded VAE object, straight into VAEDecode.
    • Steps / CFG / sampler_name / schedulerKSampler.

    For the KSampler and CheckpointLoaderSimple connections, right-click → Convert Widget to Input on the target widgets first - the README shows exactly this for ckpt_name.

    Install

    Via ComfyUI Manager (search "ComfyUI_mittimiLoadPreset"), or manually:

    cd ComfyUI/custom_nodes
    git clone https://github.com/mittimi/ComfyUI_mittimiLoadPreset
    

    Restart ComfyUI. Only dependency is toml, no model downloads.

    Before you commit

    The README's top banner says this pack is retired in favor of ComfyUI_mittimiLoadPreset2 and won't receive updates. For a one-off workflow it's perfectly serviceable; if you're building something you'll maintain, v2 is the path the author is actually supporting.

    CategorymittimiTools

    Inputs (12)

    NameTypeDefaultDescription
    presetCOMBO3 options: (testing)preset.toml, PRESET TEMPLATE.toml, [Sample] PonyRealismSetting.toml
    PosPromptASTRING
    PosPromptCSTRING
    NegPromptASTRING
    NegPromptCSTRING
    checkpoint_nameCOMBO0 options:
    ClipNumINT-1-10–-1
    vae_nameCOMBO0 options:
    StepsINT
    CFGFLOAT
    SamplerNameCOMBO34 options: euler, euler_cfg_pp, euler_ancestral, euler_ancestral_cfg_pp, heun, heunpp2, +28
    SchedulerCOMBO9 options: normal, karras, exponential, sgm_uniform, simple, ddim_uniform, +3

    Outputs (11)

    NameTypeDescription
    POS ASTRING
    POS CSTRING
    NEG ASTRING
    NEG CSTRING
    ckpt_name
    stop_at_clip_layerINT
    vaeVAE
    StepsINT
    CFGFLOAT
    sampler_nameeuler,euler_cfg_pp,euler_ancestral,euler_ancestral_cfg_pp,heun,heunpp2,dpm_2,dpm_2_ancestral,lms,dpm_fast,dpm_adaptive,dpmpp_2s_ancestral,dpmpp_2s_ancestral_cfg_pp,dpmpp_sde,dpmpp_sde_gpu,dpmpp_2m,dpmpp_2m_cfg_pp,dpmpp_2m_sde,dpmpp_2m_sde_gpu,dpmpp_3m_sde,dpmpp_3m_sde_gpu,ddpm,lcm,ipndm,ipndm_v,deis,res_multistep,res_multistep_cfg_pp,res_multistep_ancestral,res_multistep_ancestral_cfg_pp,gradient_estimation,ddim,uni_pc,uni_pc_bh2
    schedulernormal,karras,exponential,sgm_uniform,simple,ddim_uniform,beta,linear_quadratic,kl_optimal