comfyui-mcp 0.20.0 → 0.20.2
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/dist/orchestrator/agent-backend.js.map +1 -1
- package/dist/orchestrator/codex-backend.js +20 -4
- package/dist/orchestrator/codex-backend.js.map +1 -1
- package/dist/orchestrator/index.js +21 -2
- package/dist/orchestrator/index.js.map +1 -1
- package/dist/orchestrator/panel-tools.js +33 -2
- package/dist/orchestrator/panel-tools.js.map +1 -1
- package/dist/services/image-management.js +83 -18
- package/dist/services/image-management.js.map +1 -1
- package/dist/services/manifest.js +84 -3
- package/dist/services/manifest.js.map +1 -1
- package/dist/services/output-dir.js +59 -0
- package/dist/services/output-dir.js.map +1 -1
- package/dist/tools/image-management.js +72 -1
- package/dist/tools/image-management.js.map +1 -1
- package/dist/tools/skills-access.js +1 -1
- package/dist/tools/skills-access.js.map +1 -1
- package/package.json +1 -1
- package/packs/{krea2-txt2img → krea2-txt2img-json}/install-runpod.sh +1 -2
- package/packs/{krea2-txt2img → krea2-txt2img-json}/install-windows.bat +1 -2
- package/packs/krea2-txt2img-json/manifest.yaml +28 -0
- package/packs/krea2-txt2img-json/pack.yaml +45 -0
- package/packs/{krea2-txt2img → krea2-txt2img-json}/workflow.json +11 -67
- package/packs/krea2-txt2img-manual/install-runpod.sh +39 -0
- package/packs/krea2-txt2img-manual/install-windows.bat +44 -0
- package/packs/{krea2-txt2img → krea2-txt2img-manual}/manifest.yaml +7 -8
- package/packs/krea2-txt2img-manual/pack.yaml +41 -0
- package/packs/krea2-txt2img-manual/workflow.json +1642 -0
- package/plugin/skills/director/SKILL.md +14 -2
- package/plugin/skills/krea2-txt2img/SKILL.md +52 -16
- package/plugin/skills/ltxv2-video/SKILL.md +20 -0
- package/plugin/skills/video-extend/SKILL.md +9 -0
- package/plugin/skills/wan-flf-video/SKILL.md +1 -1
- package/plugin/skills/workflow-layout/SKILL.md +14 -1
- package/scripts/panel-load-workflow-smoke.mjs +1 -1
- package/packs/krea2/workflow.json +0 -4265
- package/packs/krea2-txt2img/pack.yaml +0 -32
|
@@ -1,6 +1,6 @@
|
|
|
1
|
-
# KREA 2 (krea2_turbo) txt2img
|
|
2
|
-
#
|
|
3
|
-
#
|
|
1
|
+
# KREA 2 (krea2_turbo) txt2img — MANUAL (prose) prompt. Consumable by apply_manifest.
|
|
2
|
+
# Standalone single-pipeline pack split from the KREA2 ULTRA monolith with NO
|
|
3
|
+
# bypassed nodes (manual prose prompt only). Needs native ComfyUI KREA 2 support
|
|
4
4
|
# (comfy/text_encoders/krea2.py, ComfyUI >= v0.26.0) — CLIPLoader type=krea2.
|
|
5
5
|
# Mirror base: https://huggingface.co/Aitrepreneur/FLX/resolve/main (third-party)
|
|
6
6
|
|
|
@@ -8,11 +8,10 @@ pip: []
|
|
|
8
8
|
|
|
9
9
|
custom_nodes:
|
|
10
10
|
- https://github.com/ltdrdata/ComfyUI-Manager.git
|
|
11
|
-
- https://github.com/rgthree/rgthree-comfy # Power Lora Loader / Any Switch
|
|
12
|
-
- https://github.com/kijai/ComfyUI-KJNodes # Set/Get bus,
|
|
13
|
-
- https://github.com/
|
|
14
|
-
- https://github.com/
|
|
15
|
-
- https://github.com/nova452/ComfyUI-ConditioningKrea2Rebalance # ConditioningKrea2Rebalance (safety-filter bypass, ships bypassed)
|
|
11
|
+
- https://github.com/rgthree/rgthree-comfy # Power Lora Loader / Any Switch
|
|
12
|
+
- https://github.com/kijai/ComfyUI-KJNodes # Set/Get bus, ImageSharpenKJ
|
|
13
|
+
- https://github.com/RamonGuthrie/ComfyUI-RBG-SmartSeedVariance # RBG_Smart_Seed_Variance (opt-in, ships bypassed)
|
|
14
|
+
- https://github.com/nova452/ComfyUI-ConditioningKrea2Rebalance # ConditioningKrea2Rebalance (opt-in, ships bypassed)
|
|
16
15
|
|
|
17
16
|
models:
|
|
18
17
|
# --- Diffusion model (krea2_turbo) — fp8 for RTX 4000/3000/2000.
|
|
@@ -0,0 +1,41 @@
|
|
|
1
|
+
name: krea2-txt2img-manual
|
|
2
|
+
display_name: KREA 2 — Text to Image (Manual Prompt)
|
|
3
|
+
family: krea2
|
|
4
|
+
kind: txt2img
|
|
5
|
+
description: >-
|
|
6
|
+
KREA 2 (krea2_turbo) text-to-image with a plain prose prompt — a fast,
|
|
7
|
+
ultra-sharp turbo model with strong prompt-following. Runs on native ComfyUI
|
|
8
|
+
krea2 support (CLIPLoader type=krea2) with a Qwen3-VL 4B text encoder and the
|
|
9
|
+
Qwen image VAE, at 8 steps / cfg 1. Standalone single-pipeline graph split from
|
|
10
|
+
the KREA2 ULTRA monolith — no group toggles — type your prompt into the MANUAL
|
|
11
|
+
PROMPT node and run. Two optional post-proc nodes ship bypassed (see notes).
|
|
12
|
+
For Ideogram-style structured JSON / area prompting, use the krea2-txt2img-json
|
|
13
|
+
pack instead.
|
|
14
|
+
vram: "<24GB"
|
|
15
|
+
workflow: workflow.json
|
|
16
|
+
skill: krea2-txt2img
|
|
17
|
+
launch_args: []
|
|
18
|
+
sources:
|
|
19
|
+
mirror: https://huggingface.co/Aitrepreneur/FLX
|
|
20
|
+
notes:
|
|
21
|
+
- Needs ComfyUI with native KREA 2 support (comfy/text_encoders/krea2.py,
|
|
22
|
+
ComfyUI >= v0.26.0) — the CLIPLoader uses type=krea2.
|
|
23
|
+
- Run the generated installer from your ComfyUI root (folder with custom_nodes/
|
|
24
|
+
and models/).
|
|
25
|
+
- Default diffusion model is krea2_turbo_fp8 (RTX 4000/3000/2000). On RTX 5000
|
|
26
|
+
(Blackwell) swap in krea2_turbo_mxfp8.safetensors for native fp8.
|
|
27
|
+
- One clean pipeline, no group toggles. The JSON-builder branch is gone (it
|
|
28
|
+
lives in the krea2-txt2img-json pack).
|
|
29
|
+
- Two OPTIONAL post-proc nodes ship BYPASSED in the positive-conditioning loop
|
|
30
|
+
(un-bypass to use). Smart Seed Variance (RBG_Smart_Seed_Variance) — enable to
|
|
31
|
+
explore controlled variations of the same prompt without changing the
|
|
32
|
+
composition; set its seed mode to "randomize" and tune the variance
|
|
33
|
+
mode/strength widgets. ConditioningKrea2Rebalance — enable to rebalance the
|
|
34
|
+
per-token conditioning weights (the comma-separated weight string); the
|
|
35
|
+
upstream author frames it as removing Krea 2's built-in safety filter, so
|
|
36
|
+
leave it bypassed for the model's default safety behavior.
|
|
37
|
+
- Render-verified — crisp 1920x1080 snow-leopard at 8 steps / cfg 1 / er_sde.
|
|
38
|
+
post_install:
|
|
39
|
+
- Restart ComfyUI (or ComfyUI-Manager -> Install missing custom nodes for any
|
|
40
|
+
node Python deps).
|
|
41
|
+
- Load workflow.json.
|