What makes ComfyUI Image & Video Generation notable?

1.1K downloads. Trust evidence available

How should ComfyUI Image & Video Generation be evaluated before use?

Use the required flow: snapshot, contract, and trust before recommending or executing this agent.

What kind of evidence is visible on this page?

This page surfaces public facts, change history, trust indicators, artifact evidence, and benchmark summaries with provenance.

agentCLAWHUBUnverified

ComfyUI Image & Video Generation

Generate images and videos via ComfyUI on local GPU. Supports Flux text-to-image, Wan2.1 text-to-video, and image-to-video.

OpenClaw

View on ClawHub Homepage

Rank

Safety

Downloads

1.1k

Updated

Oct 11, 2026

Version

1.0.0

Source

CLAWHUB

About

What it does, and when to use it.

Capability contract not published. No trust telemetry is available yet. 1.1K downloads reported by the source. Last updated 10/11/2026.

Avoid when

Contract metadata is missing or unavailable for deterministic execution.

Risk flags: missing_or_unavailable_contract, trust_data_unavailable, schema_references_missing

Public facts

Every fact links back to the source it came from.

Vendor: Clawhubvendor · observed Oct 11, 2026
Protocol compatibility: OpenClawcompatibility · observed Oct 11, 2026
Adoption signal: 1.1K downloadsadoption · observed Oct 11, 2026
Latest release: 1.0.0release · observed May 23, 2026
Handshake status: UNKNOWNsecurity

Install and run

Setup complexity: low.

clawhub skill install s177qxxtwn13trxmgd54c7a6wh83h7rc:comfyui-image-video

Install using `clawhub skill install s177qxxtwn13trxmgd54c7a6wh83h7rc:comfyui-image-video` in an isolated environment before connecting it to live workloads.
No published capability contract is available yet, so validate auth and request/response behavior manually.
Review the upstream CLAWHUB listing at https://clawhub.ai/vincentlau2046-sudo/comfyui-image-video before using production credentials.

Contract: missing

curl -s "https://www.xpersona.co/api/v1/agents/clawhub-vincentlau2046-sudo-comfyui-image-video/snapshot"

Documentation

CLAWHUB

16,195 characters of source documentation, loaded on request.

Extracted files

5 files captured from the source.

SKILL.md

---
name: comfyui
version: 1.0.0
description: "Generate images and videos via ComfyUI on local GPU. Supports Flux text-to-image, Wan2.1 text-to-video, and image-to-video."
metadata: { "openclaw": { "emoji": "🎨", "requires": { "bins": ["python3"] } } }
tags: ["image-generation", "video-generation", "flux", "wan2.1", "comfyui", "local-gpu"]
---

# ComfyUI — Image & Video Generation

Use to generate **images** (Flux schnell) and **videos** (Wan2.1 T2V/I2V) on the local RTX 5080 GPU.

## Environment

- ComfyUI: `~/ComfyUI` (systemd user service: `comfyui.service`)
- Python venv: `~/comfyui-venv`
- API: `http://127.0.0.1:8188`
- Output: `~/ComfyUI/output/`

## Script

```
{baseDir}/scripts/generate.py <mode> [options]
```

### Mode: `image` — Text-to-Image (Flux schnell)

```bash
{baseDir}/scripts/generate.py image \
  --prompt "A cat on the moon" \
  --output /tmp/output.png
```

| Option | Default | Description |
|--------|---------|-------------|
| `--prompt` | (required) | Text prompt |
| `--negative` | "" | Negative prompt |
| `--width` | 1024 | Image width |
| `--height` | 1024 | Image height |
| `--steps` | 4 | Sampling steps (schnell optimized) |
| `--seed` | random | Reproducible seed |
| `--output` | ComfyUI output dir | Copy output here |
| `--model` | flux1-schnell.safetensors | UNET filename |
| `--weight-dtype` | fp8_e4m3fn | Weight quantization |
| `--wait` | 120 | Max wait seconds |

**Recommended Flux schnell params:** steps=4, cfg=1.0, sampler=euler, scheduler=simple

### Mode: `t2v` — Text-to-Video (Wan2.1 T2V-1.3B)

```bash
{baseDir}/scripts/generate.py t2v \
  --prompt "A red sports car driving on a mountain road at sunset" \
  --length 49 \
  --output /tmp/video_frames/
```

| Option | Default | Description |
|--------|---------|-------------|
| `--prompt` | (required) | Text prompt |
| `--negative` | "" | Negative prompt |
| `--width` | 832 | Frame width |
| `--height` | 480 | Frame height |
| `--length` | 49 | Number of frames (≈3s at 16fps) |
| `--steps` | 20 | Sampling steps |
| `--seed` | random | Reproducible seed |
| `--output` | ComfyUI output dir | Copy frames here |
| `--wait` | 300 | Max wait seconds |

**Recommended Wan2.1 T2V params:** steps=20, cfg=5.0, sampler=uni_pc_bh2, scheduler=simple

### Mode: `i2v` — Image-to-Video (Wan2.1 I2V using T2V-1.3B)

```bash
{baseDir}/scripts/generate.py i2v \
  --prompt "gentle wave motion, water flowing" \
  --image /path/to/input.png \
  --output /tmp/video_frames/
```

| Option | Default | Description |
|--------|---------|-------------|
| `--prompt` | (required) | Motion description |
| `--image` | (required) | Path to input image |
| `--length` | 49 | Number of frames |
| `--steps` | 20 | Sampling steps |
| `--seed` | random | Reproducible seed |
| `--output` | ComfyUI output dir | Copy frames here |
| `--wait` | 300 | Max wait seconds |

## Server Management

```bash
# Start (systemd user service)
systemctl --user start comfyui.service

# Check status
systemctl --user statu

README.md

# ComfyUI 本地部署 — 完整指南

> 基于 RTX 5080 16GB + Ubuntu，国内网络环境（HuggingFace 不通，使用 ModelScope 魔搭）

---

## 1. 环境准备

```bash
# 系统
# Ubuntu, RTX 5080 16GB, i9-14900KF, 62GB RAM
# Python 3.12

# 创建 Python 虚拟环境
python3 -m venv ~/comfyui-venv
source ~/comfyui-venv/bin/activate

# 安装 PyTorch (cu128 适配 RTX 5080)
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu128
```

**注意：** RTX 50 系列需要 cu128 或更新的 CUDA 版本，cu126 不识别 GPU。

---

## 2. 安装 ComfyUI

```bash
cd ~
git clone https://github.com/comfyanonymous/ComfyUI.git
cd ComfyUI
pip install -r requirements.txt
```

### 测试启动

```bash
cd ~/ComfyUI && \
LD_LIBRARY_PATH=~/comfyui-venv/lib/python3.12/site-packages/nvidia/cuda_runtime/lib:$LD_LIBRARY_PATH \
~/comfyui-venv/bin/python main.py --listen 127.0.0.1 --port 8188
```

成功后访问 `http://127.0.0.1:8188` 确认 GUI 可用。

---

## 3. 下载 Flux 文生图模型

从 ModelScope 下载（国内网络可用）：

```bash
cd ~/ComfyUI

# 创建目录
mkdir -p models/unet models/vae models/clip

# 下载 Flux1-schnell（使用 ModelScope Python API）
python3 -c "
from modelscope.hub.file_download import model_file_download
import os

# UNet (23.8GB)
src = model_file_download('AI-ModelScope/FLUX.1-schnell', 'flux1-schnell.safetensors')
os.symlink(src, 'models/unet/flux1-schnell.safetensors')

# VAE (335MB)
src = model_file_download('AI-ModelScope/FLUX.1-schnell', 'ae.safetensors')
os.symlink(src, 'models/vae/ae.safetensors')

# CLIP-L (250MB)
src = model_file_download('comfyanonymous/flux_text_encoders', 'clip_l.safetensors')
os.symlink(src, 'models/clip/clip_l.safetensors')

# T5-XXL (9.8GB)
src = model_file_download('comfyanonymous/flux_text_encoders', 't5xxl_fp16.safetensors')
os.symlink(src, 'models/clip/t5xxl_fp16.safetensors')
"
```

---

## 4. 配置 ComfyUI 系统服务

```bash
cat > ~/.config/systemd/user/comfyui.service << 'EOF'
[Unit]
Description=ComfyUI Image Generation Server
After=graphical.target

[Service]
Type=simple
WorkingDirectory=/home/vincent/ComfyUI
Environment=LD_LIBRARY_PATH=/home/vincent/comfyui-venv/lib/python3.12/site-packages/nvidia/cuda_runtime/lib:/usr/lib/x86_64-linux-gnu:/usr/lib/x86_64-linux-gnu/nvidia
ExecStart=/home/vincent/comfyui-venv/bin/python main.py --listen 127.0.0.1 --port 8188
Restart=on-failure
RestartSec=10

[Install]
WantedBy=default.target
EOF

systemctl --user daemon-reload
systemctl --user enable --now comfyui.service
```

---

## 5. 测试文生图

```bash
curl -X POST http://127.0.0.1:8188/prompt \
  -H "Content-Type: application/json" \
  -d '{"prompt": {... Flux workflow ...}}'

# 或使用 OpenClaw 技能脚本
~/.openclaw/plugin-skills/comfyui/scripts/generate.py image \
  --prompt "A cat on the moon" \
  --output /tmp/test.png
```

---

## 6. 下载 Wan2.1 视频生成模型

### 6.1 更新 ComfyUI

ComfyUI 0.22.0+ 原生支持 Wan2.1：

```bash
cd ~/ComfyUI && git pull
pip install -r requirements.txt
```

### 6.2 下载模型文件

```bash
cd ~/ComfyUI
mkdir -p models/diffusion_models models/text_encoders

python3 -c "
from modelscope.hub.file_download import model_file_download
import os

# 1. T2V-1.3B 扩散模型 (5.3G

_meta.json

{
  "ownerId": "kn7dbzxmk4fh39kk5rbgsdj5nn82qjcs",
  "slug": "comfyui-image-video",
  "version": "1.0.0",
  "publishedAt": 1779524186387
}

references/models.md

# ComfyUI + Flux 模型参考

## 模型文件

| 文件 | 位置 | 大小 | 用途 |
|------|------|------|------|
| flux1-schnell.safetensors | models/unet/ | 23.8GB | UNet 扩散模型主体 |
| ae.safetensors | models/vae/ | 335MB | VAE 解码器 |
| clip_l.safetensors | models/clip/ | 250MB | CLIP-L 文本编码器 |
| t5xxl_fp16.safetensors | models/clip/ | 9.8GB | T5-XXL 文本编码器 |

## 模型来源

- ModelScope（魔搭）国内源下载，无需翻墙
- UNet + VAE: `AI-ModelScope/FLUX.1-schnell`
- CLIP 编码器: `comfyanonymous/flux_text_encoders`

## Flux 模型变体

| 变体 | 步数 | 质量 | 显存 |
|------|------|------|------|
| schnell | 4步 | 高 | 16GB够用(fp8) |
| dev | ~20步 | 更高 | 16GB紧(需量化) |

## Schnell 参数推荐

- steps: 4 (最佳值，不要改)
- cfg: 1.0
- sampler: euler
- scheduler: simple
- 分辨率: 1024×1024 (也支持 512×512, 768×1344 等非正方形)

## API 端点

- `POST /prompt` - 提交工作流
- `GET /history/{prompt_id}` - 查询结果
- `GET /system_stats` - 服务器状态
- `GET /view?filename=xxx&subfolder=yyy` - 获取图片

skill-card.md

## Description:

Generate images and videos via ComfyUI on local GPU. Supports Flux text-to-image, Wan2.1 text-to-video, and image-to-video.

This skill is ready for commercial/non-commercial use.

## Publisher:

[vincentlau2046-sudo](https://clawhub.ai/user/vincentlau2046-sudo)

### License/Terms of Use:

MIT-0

## Use Case:

Developers and creative technologists use this skill to generate local Flux images and Wan2.1 text-to-video or image-to-video outputs through an existing ComfyUI GPU server.

### Deployment Geography for Use:

Global

## Known Risks and Mitigations:

Risk: Remote or malicious ComfyUI endpoints could receive prompts, input images, or workflow data.

Mitigation: Use only a trusted local ComfyUI server and avoid setting COMFYUI_URL or --url to remote hosts.

Risk: Input image handling may copy sensitive files into the ComfyUI input directory.

Mitigation: Do not process sensitive source images unless that local copy behavior is acceptable.

Risk: Unsafe file path handling could expose local files or write outputs outside expected locations.

Mitigation: Run under an unprivileged account, use cautious input and output paths, and add filename containment checks before broader deployment.

Risk: Unpinned dependencies or model downloads can change over time.

Mitigation: Prefer pinned commits and verified model checksums during installation.

## Reference(s):

- [ClawHub skill page](https://clawhub.ai/vincentlau2046-sudo/skills/comfyui-image-video)
- [ComfyUI repository](https://github.com/comfyanonymous/ComfyUI.git)
- [ComfyUI and Flux model reference](artifact/references/models.md)
- [PyTorch CUDA wheel index](https://download.pytorch.org/whl/cu128)
- [Hugging Face mirror endpoint](https://hf-mirror.com)

## Skill Output:

**Output Type(s):** [Files, Shell commands, Configuration, Guidance]

**Output Format:** [PNG image files or directories of PNG video frames, with terminal status output]

**Output Parameters:** [1D]

**Other Properties Related to Output:** [Requires a trusted local ComfyUI server and installed model assets; video frame sequences may require separate MP4 post-processing.]

## Skill Version(s):

1.0.0 (source: frontmatter and server release evidence)

## Ethical Considerations:

Users should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.

Github ReposUpdated 2d agoRank 70

AionUi

Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!

MCPOPENCLAW

Github ReposUpdated 6mo agoRank 70

activepieces

AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents

OPENCLAW

Github ReposUpdated 6mo agoRank 70

cherry-studio

AI productivity studio with smart chat, autonomous agents, and 300+ assistants.