museav-cli 2.4.0 → 2.4.1
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/AGENTS.md +6 -0
- package/README.md +23 -0
- package/package.json +1 -1
package/AGENTS.md
CHANGED
|
@@ -64,8 +64,14 @@ museav upload ./face.png
|
|
|
64
64
|
# compress: resize/re-encode via sharp. Output <name>-min.<fmt>, never overwrites input.
|
|
65
65
|
# remove-bg: ISNet via onnxruntime-node → alpha PNG (~170MB model auto-downloaded
|
|
66
66
|
# to ~/.museav-models on first use; %USERPROFILE%\.museav-models on Windows).
|
|
67
|
+
# upscale: Real-ESRGAN via Vulkan GPU → 2M image becomes 10M+ (~65MB engine+model
|
|
68
|
+
# auto-downloaded once to ~/.museav-bin + ~/.museav-models).
|
|
69
|
+
# remove-watermark: pixel-heuristic auto-locate (zero model deps) + LaMa inpainting
|
|
70
|
+
# → clean PNG (~200MB model on-demand, freed after run).
|
|
67
71
|
museav compress ./photo.jpg --max-edge 800 --format webp --quality 70
|
|
68
72
|
museav remove-bg ./shoe.png # stdout: path to <name>-nobg.png
|
|
73
|
+
museav upscale ./photo.jpg --scale 4 # stdout: path to <name>-4x.png
|
|
74
|
+
museav remove-watermark ./poster.jpg # stdout: path to <name>-clean.png
|
|
69
75
|
|
|
70
76
|
# List your own (or, if using a tenant apiKey, your tenant's) recent jobs as JSON
|
|
71
77
|
museav jobs --limit 10 --status failed
|
package/README.md
CHANGED
|
@@ -85,6 +85,27 @@
|
|
|
85
85
|
|
|
86
86
|
两条路径互不依赖,选跟你身份匹配的那条,不用两个都配。下面「快速开始」走的是平台用户(login)这条路;如果你是租户,直接跳到下面「B 端 / CI 场景」那一节。
|
|
87
87
|
|
|
88
|
+
## 能力速览(小白先看这个)
|
|
89
|
+
|
|
90
|
+
`museav` 能干什么,一句话一组:
|
|
91
|
+
|
|
92
|
+
| 分类 | 能力 | 命令 |
|
|
93
|
+
|---|---|---|
|
|
94
|
+
| **AI 生成**(中台) | 文字出图 / 垫图出图 / 技能出图 / 模板出图 | `gen`(含 `--ref` 垫图、`--transparent` 透明底) |
|
|
95
|
+
| | 文生视频 / 图生视频 | `gen --video` |
|
|
96
|
+
| | 从已有图片反推提示词 | `reverse` |
|
|
97
|
+
| | 一张图做成可复用模板 | `image-to-template` |
|
|
98
|
+
| | 建/查图片模板、视频模板 | `templates` / `video-templates` |
|
|
99
|
+
| **项目管理**(中台) | 工作区(项目)+ 每个项目自己的素材库 | `projects`、`projects assets` |
|
|
100
|
+
| | 素材直链垫图、作品归档进项目 | `gen --project`、`jobs --project` |
|
|
101
|
+
| **本地图像处理**(免登录/免费/离线) | 压缩图片 | `compress` |
|
|
102
|
+
| | 抠图去背景(输出透明 PNG) | `remove-bg` |
|
|
103
|
+
| | **放大清晰度**(2M → 10M+ 级) | `upscale` |
|
|
104
|
+
| | 去水印 | `remove-watermark` |
|
|
105
|
+
| **素材与统计** | 上传素材 / 查任务 / 查余额 / 查模型 / 查身份 | `upload` / `jobs` / `balance` / `models` / `whoami` |
|
|
106
|
+
|
|
107
|
+
> 本地工具(`compress` / `remove-bg` / `upscale` / `remove-watermark`)**不用登录、不花一分钱**,装了就能用;其余命令需要一个凭证(个人 `login` 或租户 apiKey)。
|
|
108
|
+
|
|
88
109
|
---
|
|
89
110
|
|
|
90
111
|
## 快速开始
|
|
@@ -476,6 +497,8 @@ console.log(r.sculpt.light) // 光影分析
|
|
|
476
497
|
| `upload <file>` | 上传素材(图片/音频/视频) | 公网直链 |
|
|
477
498
|
| `compress <file>` | 本地压缩图片(免登录,`--max-edge`/`--quality`/`--format`) | 产物路径 |
|
|
478
499
|
| `remove-bg <file>` | 本地抠图去背景(免登录,输出带 alpha 的 PNG) | 产物路径 |
|
|
500
|
+
| `upscale <file>` | 本地超分放大,2M 图变 10M+(免登录,Real-ESRGAN GPU,`--scale 2/3/4`) | 产物路径 |
|
|
501
|
+
| `remove-watermark <file>` | 本地去水印(免登录,自动定位 + LaMa 修复,`--mask` 手工兜底) | 产物路径 |
|
|
479
502
|
| `projects` | 工作区(项目)列表:一账户多项目,各带自己的素材库 | 项目 id 列表 |
|
|
480
503
|
| `projects assets --project` | 项目素材库:ls / add / rm(垫图母版,人像库/产品库各管各的) | `id<TAB>url` 行 |
|
|
481
504
|
| `models` | 可用模型 | 模型名列表 |
|