According to AlphaMoat, as of 2026-09, jimeng generator has 2,176 downloads and 0 stars — ranked #4,307 of 5,078 AI skills overall, and #431 of 512 in Design & Media.
| Platform | Installs | Downloads | Stars | Comments |
|---|---|---|---|---|
| ClawHub ↗ | 32 | 1.1K | - | - |
| SkillHub ↗ | 539 | 2.2K | - | — |
Platforms count differently and update at different times.
It renders images from text, accepts multiple reference images for editing, and adapts the aspect ratio when one is described in the prompt. Before use, put an Access Key and Secret Key into a project-level env file, or switch to an AK plus token pair for temporary STS credentials. After installing dependencies, the script submits a task and waits for the saved result. Review the destination/account and data sharing before use; confirm consequential external actions and independently verify generated or retrieved results.
Profile published 2026-10-27
Generated from the skill’s documentation as examples; not results of actual runs.
The counting method changed in 2026-09: skills with the same name are now counted separately by author, and the full catalog is covered. Dashed lines show months before the change; they are not compared with later months, and no growth is computed across the change.
| Month | Downloads | MoM | Stars |
|---|---|---|---|
| 2026-09 | 2.2K | — | - |
| 2026-08 | 1.8K | +5.4% | - |
| 2026-07 | 1.7K | +10.5% | - |
| 2026-06 | 1.5K | +6.1% | - |
| 2026-05 | 1.4K | — | - |
| 2026-03 | 882 | — | - |
The author’s original file, shown as plain text without edits or review by AlphaMoat. Links are not clickable; check commands before running them. Fetched 2026-10-23.
--- name: jimeng-generator description: 即梦 4.0 图片生成器,通过文本描述生成高质量图片,支持多图编辑与智能比例。 --- # 即梦 4.0 图片生成器 基于火山引擎即梦 AI 4.0 生成图片。一条命令完成:提交任务 → 等待完成 → 保存图片。 ## 配置凭证
使用前需要配置火山引擎凭证。请按以下步骤操作:
1. 前往 [火山引擎控制台](https://console.volcengine.com/) → 访问控制 → 密钥管理,获取 Access Key 和 Secret Key
2. 在项目根目录创建 `.env` 文件,写入你的凭证:
```bash
VOLCENGINE_AK=你的AccessKey
VOLCENGINE_SK=你的SecretKey
```
> 如果使用 STS 临时凭证,改为填写 AK + Token:
> ```bash
> VOLCENGINE_AK=你的AccessKey
> VOLCENGINE_TOKEN=你的SecurityToken
> ```
项目已提供 `.env.example` 模板,也可以直接复制后修改:
```bash
cp .env.example .env
```
`.env` 文件已被 `.gitignore` 忽略,不会提交到仓库,请放心填写真实凭证。
## 安装
```bash
npm install
```
## 基本用法
```bash
npx ts-node scripts/generate.ts "提示词"
```
脚本会自动提交 → 轮询 → 保存图片到 `./output/`。
## 完整用法
```bash
npx ts-node scripts/generate.ts "提示词" [选项]
```
### 选项
| 选项 | 说明 | 默认值 |
|------|------|--------|
| `--images <url,...>` | 参考图片 URL,逗号分隔,最多 10 张 | — |
| `--width <n>` | 输出宽度 | 自动 |
| `--height <n>` | 输出高度 | 自动 |
| `--size <n>` | 输出面积 | 自动 |
| `--scale <0-1>` | 文本影响程度 | 0.5 |
| `--single` | 强制单图输出 | false |
| `--out <dir>` | 输出目录 | ./output |
| `--no-save` | 不保存,只输出 URL | false |
| `--interval <ms>` | 轮询间隔 | 3000 |
| `--timeout <ms>` | 最大等待时间 | 180000 |
| `--debug` | 调试模式 | false |
## 使用示例
### 文生图
```bash
npx ts-node scripts/generate.ts "水墨山水画"
```
### 指定尺寸
```bash
npx ts-node scripts/generate.ts "赛博朋克城市" --width 2560 --height 1440
```
### 图片编辑
```bash
npx ts-node scripts/generate.ts "背景换成星空" --images "https://example.com/photo.jpg"
```
### 多图组合
```bash
npx ts-node scripts/generate.ts "合成一张合照" --images "https://a.jpg,https://b.jpg"
```
### 强制单图 + 高影响
```bash
npx ts-node scripts/generate.ts "精细插画风格的城堡" --single --scale 0.8
```
## 输出格式
脚本结果通过 **stdout** 输出一行 JSON,便于解析。**请用其中的 `files` 数组直接作为“生成结果”返回给用户**(本地文件路径,可用来展示或附件),不要只贴直链或把本地路径藏在文案里。
成功时示例:
```json
{
"success": true,
"taskId": "7392616336519610409",
"prompt": "水墨山水画",
"count": 1,
"files": ["/path/to/output/1.png"],
"urls": ["https://..."]
}
```
| 字段 | 说明 |
|------|------|
| `files` | **本地文件路径数组**,即生成图片的保存位置,应作为主结果直接返回给用户 |
| `urls` | 图片直链(可选),仅作备用 |
| `taskId` | 任务 ID |
| `count` | 图片张数 |
集成建议:解析 stdout 该行 JSON,若 `success === true`,则把 `files` 中的路径作为“已生成的图片”直接提供给用户(例如展示图片或作为附件),无需再组织成“直链在这:… 本地文件在:…”的文案。
失败时错误信息在 stderr,JSON 形如:
```json
{
"success": false,
"error": { "code": "FAILED", "message": "错误描述" }
}
```
## 即梦 4.0 特性
- **智能比例**:可在 prompt 中描述比例,模型自动适配最优宽高
- **多图输入**:最多 10 张参考图,支持图片编辑和多图组合
- **多图输出**:单次最多输出 15 张关联图片
- **4K 输出**:支持从 1K 到 4K 分辨率
- **中文增强**:显著提升中文生成准确率
## 推荐尺寸
| 分辨率 | 1:1 | 4:3 | 3:2 | 16:9 | 21:9 |
|--------|-----|-----|-----|------|------|
| 2K | 2048×2048 | 2304×1728 | 2496×1664 | 2560×1440 | 3024×1296 |
| 4K | 4096×4096 | 4694×3520 | 4992×3328 | 5404×3040 | 6198×2656 |
ContextWeave one-shot architecture diagram generation
A script-based skill that generates or edits images with Nano Banana Pro.
Runs the whisper command line to transcribe audio locally.
A skill that generates and edits diagrams through an MCP server.
Design intelligence and implementation guidance for polished interfaces, from layout and flows to tokens and frontend changes.
An AI video generation and orchestration skill across several model APIs.
Pull a frame out of a video with ffmpeg, or produce a thumbnail for a quick look.
Turn audio files into text with local Whisper, running fully offline once the model is downloaded.
Data month: 2026-09 · Downloads, stars and installs are aggregated monthly from public skill registries (ClawHub, SkillHub). See methodology.