## TT FLUX Workflow Settings (Advanced) *TT_FluxWorkflowSettingsAdvancedNode* *This description was generated with AI assistance. If you spot any errors, please create an issue on GitHub.* Extended FLUX workflow settings node with built-in prompt inputs for CLIP-L and T5-XXL encoders. Combines sampler configuration and text prompts in a single node. ### Inputs | Parameter | Type | Required | Description | |-------------------|--------|----------|--------------------------------------| | `seed` | INT | Yes | Generation seed | | `steps` | INT | Yes | Number of sampling steps, default 30 | | `cfg` | FLOAT | Yes | Guidance scale, default 5.0 | | `sampler_name` | COMBO | Yes | Sampler algorithm | | `scheduler` | COMBO | Yes | Noise schedule | | `clip_l_positive` | STRING | Yes | CLIP-L positive prompt | | `t5xxl_positive` | STRING | Yes | T5-XXL positive prompt | | `clip_l_negative` | STRING | Yes | CLIP-L negative prompt | | `t5xxl_negative` | STRING | Yes | T5-XXL negative prompt | | `lora_triggers` | STRING | No | LoRA trigger words (advanced) | | `guidance` | FLOAT | Yes | FLUX guidance scale, default 3.0 | ### Outputs | Parameter | Type | Description | |-------------------|-----------------|------------------------------------------| | `WORKFLOW_CONFIG` | WORKFLOW_CONFIG | Workflow config object for context nodes | | `SEED` | INT | Generation seed | | `STEPS` | INT | Number of sampling steps | | `CFG` | FLOAT | Guidance scale | | `SAMPLER_NAME` | STRING | Sampler algorithm name | | `SCHEDULER` | STRING | Noise schedule name | | `CLIP_L_POSITIVE` | STRING | CLIP-L positive prompt | | `T5XXL_POSITIVE` | STRING | T5-XXL positive prompt | | `CLIP_L_NEGATIVE` | STRING | CLIP-L negative prompt | | `T5XXL_NEGATIVE` | STRING | T5-XXL negative prompt | | `LORA_TRIGGERS` | STRING | LoRA trigger words | | `GUIDANCE` | FLOAT | FLUX guidance scale | ### Usage Use when you want prompts and sampler settings in one place. Connect `WORKFLOW_CONFIG` to context nodes, or use individual outputs for direct connections. LoRA trigger words are appended to the positive prompt automatically in the text encoder node. --- [View on GitHub](https://github.com/tenser-tensor/ComfyUI-TenserTensor)