Add second workflow batch for Z-Image, LTX, and MiniMax

models/ltx/ltx23-character-sheet-best-face-id/README.md

@@ -0,0 +1,41 @@
+# LTX 2.3 Character Sheet Best Face ID
+
+## Summary
+
+`LTX 2.3` character-sheet / face-reference workflow using `Best-Face-ID`-style LoRA support.
+
+## Model Family
+
+- `ltx`
+
+## Status
+
+- `imported-only`
+
+## Source
+
+- Workflow pack page: <https://huggingface.co/RuneXX/LTX-2.3-Workflows>
+- Workflow file: <https://huggingface.co/RuneXX/LTX-2.3-Workflows/resolve/main/Multi-ref-character-sheet/LTX-2.3_-_T2V_R2V_character-sheet_Best-Face-ID-Lora.json?download=true>
+- Original publisher: `RuneXX`
+- Date imported: `2026-08-08`
+
+## Developer notes
+
+- Comes from the multi-reference / character-sheet section of the RuneXX pack.
+- Uses more identity/reference-specific conditioning nodes than the simple LTX workflows.
+
+## Our notes
+
+- This is exactly the kind of workflow worth collecting if character consistency matters.
+- Likely too heavy to be the very first `3070` test, but high-value for later comparison and optimization.
+- Strong input candidate for a future `optimized/image-to-video/` or consistency-oriented LTX master.
+
+## Required custom nodes
+
+- `ComfyUI-KJNodes`
+- `rgthree-comfy`
+- identity/reference helper nodes used by the workflow pack
+
+## Notes
+
+- Best current LTX character-consistency import in the repo.

models/ltx/ltx23-first-last-frame/README.md

@@ -0,0 +1,40 @@
+# LTX 2.3 First Last Frame
+
+## Summary
+
+`LTX 2.3` first-last-frame workflow for controlled transition video generation.
+
+## Model Family
+
+- `ltx`
+
+## Status
+
+- `imported-only`
+
+## Source
+
+- Workflow pack page: <https://huggingface.co/RuneXX/LTX-2.3-Workflows>
+- Workflow file: <https://huggingface.co/RuneXX/LTX-2.3-Workflows/resolve/main/First-Last-Frame/LTX-2.3_-_FLF2V_First-Last-Frame.json?download=true>
+- Original publisher: `RuneXX`
+- Date imported: `2026-08-08`
+
+## Developer notes
+
+- Part of the RuneXX extracted-model workflow pack.
+- Includes several attention/sage/memory-efficiency helpers compared with the simpler LTX graphs.
+
+## Our notes
+
+- This is one of the most relevant `LTX 2.3` workflows for you because first/last-frame control is a practical, non-toy use case.
+- It is heavier than the simple single-pass workflows, so it is more of a second-step test after confirming LTX is even workable on the `3070`.
+
+## Required custom nodes
+
+- `ComfyUI-KJNodes`
+- `rgthree-comfy`
+- attention/memory helper nodes used by the RuneXX pack
+
+## Notes
+
+- Best current LTX first-last-frame import in the repo.

models/minimax/minimax-h3-int8-r2v-community/README.md

@@ -0,0 +1,54 @@
+# MiniMax H3 INT8 R2V Community
+
+## Summary
+
+Community `MiniMax H3` reference-to-video workflow built around `INT8` inference and an explicit acceleration stack.
+
+## Model Family
+
+- `minimax`
+
+## Status
+
+- `imported-only`
+
+## Source
+
+- Workflow repo: <https://huggingface.co/javawock7618/comfy-MiniMax-H3-workflows>
+- Workflow file: <https://huggingface.co/javawock7618/comfy-MiniMax-H3-workflows/resolve/main/MiniMax_int8_R2V-javano2608.5.json?download=true>
+- Original publisher: `javawock7618`
+- Date imported: `2026-08-08`
+
+## Developer notes
+
+- The repo describes this as part of an `INT8 FLF & Reference Video` collection.
+- The author explicitly targets:
+  - reduced VRAM usage
+  - faster inference
+  - improved memory efficiency
+- The documented acceleration stack includes:
+  - `Sol-Attn`
+  - `Spectrum`
+  - optional `Lightx2v LoRA`
+  - optional `MiniMax-H3-Turbo-Lora`
+  - optional latent upscaler
+
+## Our notes
+
+- This is the strongest low-VRAM-oriented MiniMax import we have so far because it is built around `INT8` from the start.
+- Still not guaranteed easy on a `3070`, but much more aligned with your constraints than generic heavyweight MiniMax examples.
+- Good candidate once you want to test whether MiniMax reference-driven video is remotely viable locally.
+
+## Required custom nodes
+
+- `ComfyUI-SolAttn_triton`
+- `ComfyUI-Spectrum-MiniMax-H3`
+- optional `ComfyUI-H3-Motion-Context` for related FLF workflows
+- optional `ComfyUI-MiniMaxH3_LatentUpscaler`
+- `Kijai/MiniMax-H3_comfy` for Lightx2v acceleration
+- optional `MiniMax-H3-Turbo-Lora-ComfyUI`
+
+## Notes
+
+- Best current MiniMax low-VRAM / INT8 reference-video workflow in the repo.
+- More of a carefully chosen experimental path than a guaranteed easy win.

models/zimage/z-image-turbo-controlnet-union-community/README.md

@@ -0,0 +1,39 @@
+# Z-Image Turbo ControlNet Union Community
+
+## Summary
+
+Community `Z-Image Turbo` workflow focused on ControlNet / union-style structural guidance.
+
+## Model Family
+
+- `zimage`
+
+## Status
+
+- `imported-only`
+
+## Source
+
+- Workflow repo: <https://huggingface.co/axiomgraph/mycomfyuiworkflows>
+- Workflow file: <https://huggingface.co/axiomgraph/mycomfyuiworkflows/resolve/main/Z%20Image%20Turbo/Z%20Image%20Turbo%20Fun%20controlnet%20union%202.1.json?download=true>
+- Original publisher: `axiomgraph`
+- Date imported: `2026-08-08`
+
+## Developer notes
+
+- Same community workflow collection as the inpainting variant.
+- Uses `ZImageFunControlnet` and a patched model path, so this is clearly aimed at guided generation rather than plain prompting.
+
+## Our notes
+
+- Worth keeping because structure-guided workflows often matter more in practice than plain T2I.
+- Good second-step `Z-Image` test after the official example, especially if you care about controlled composition.
+
+## Required custom nodes
+
+- `rgthree-comfy`
+- whatever custom package provides `ZImageFunControlnet`
+
+## Notes
+
+- Good candidate for future `optimized/text-to-image/` or `optimized/image-edit/` comparison if it behaves well.

models/zimage/z-image-turbo-inpainting-community/README.md

@@ -0,0 +1,40 @@
+# Z-Image Turbo Inpainting Community
+
+## Summary
+
+Community `Z-Image Turbo` inpainting workflow from a Hugging Face workflow repository. Good for testing whether `Z-Image Turbo` can pull useful edit work on lighter local hardware.
+
+## Model Family
+
+- `zimage`
+
+## Status
+
+- `imported-only`
+
+## Source
+
+- Workflow repo: <https://huggingface.co/axiomgraph/mycomfyuiworkflows>
+- Workflow file: <https://huggingface.co/axiomgraph/mycomfyuiworkflows/resolve/main/Z%20Image%20Turbo/Z%20Image%20Turbo%20Fun%202_1%20InPainting%204.0.json?download=true>
+- Original publisher: `axiomgraph`
+- Date imported: `2026-08-08`
+
+## Developer notes
+
+- The workflow repo is a personal collection rather than an official package.
+- This workflow uses `LanPaint_KSampler` and a `ZImageFunControlnet` node, so it is not a bare-minimum stock graph.
+
+## Our notes
+
+- Interesting because it gives us an actual `Z-Image Turbo` edit path, not just text-to-image.
+- On a `3070`, this feels like one of the most plausible community `Z-Image` workflows to test soon because editing often matters more than raw generation novelty.
+
+## Required custom nodes
+
+- `rgthree-comfy`
+- whatever custom package provides `LanPaint_KSampler`
+- whatever custom package provides `ZImageFunControlnet`
+
+## Notes
+
+- Strong candidate for the next `Z-Image` testing round.