aimlapi-media-gen

TotalClaw 作者 totalclaw

根据提示通过 AIMLAPI 生成图像或视频。当 Codex 需要具有重试、显式用户代理标头和异步视频轮询功能的可靠 AI/ML API 媒体生成时使用。

安装 / 下载方式

TotalClaw CLI推荐

totalclaw skills install shop:totalclaw:totalclaw~aimlapihello-aiml-image-video

cURL直接下载，无需登录

curl -fsSL https://skills.taituai.com/api/skills/totalclaw%3Atotalclaw~aimlapihello-aiml-image-video/file -o aimlapihello-aiml-image-video.md

# AIMLAPI Media Generation

## Overview

Generate images and videos via AIMLAPI with scripts that include retries, API key file fallback, verbose logs, and required `User-Agent` headers on every request.

## Quick start

```bash
export AIMLAPI_API_KEY="sk-aimlapi-..."
python3 {baseDir}/scripts/gen_image.py --prompt "ultra-detailed studio photo of a lobster astronaut"
python3 {baseDir}/scripts/gen_video.py --prompt "slow drone shot of a foggy forest"
```

## Tasks

### Generate images

Use `scripts/gen_image.py` with `/v1/images/generations`.

```bash
python3 {baseDir}/scripts/gen_image.py \
  --prompt "cozy cabin in a snowy forest" \
  --model aimlapi/openai/gpt-image-1 \
  --size 1024x1024 \
  --count 2 \
  --retry-max 4 \
  --user-agent "openclaw-custom/1.0" \
  --out-dir ./out/images
```

### Generate videos (async AIMLAPI flow)

Use `scripts/gen_video.py` with the real async flow:

1. `POST /v2/video/generations` (create task)
2. `GET /v2/video/generations?generation_id=...` (poll status)
3. download `video.url` when status is completed

```bash
python3 {baseDir}/scripts/gen_video.py \
  --model google/veo-3.1-t2v-fast \
  --prompt "time-lapse of clouds over a mountain range" \
  --poll-interval 10 \
  --max-wait 1000 \
  --user-agent "openclaw-custom/1.0" \
  --out-dir ./out/videos
```

## References

- `references/aimlapi-media.md`: endpoint notes, async polling statuses, and troubleshooting.
- `README.md`: changelog-style summary of new instructions.

---

## 中文说明

# AIMLAPI 媒体生成

## 概述

通过 AIMLAPI 生成图像和视频，脚本包含重试、API 密钥文件回退、详细日志，以及每次请求所需的 `User-Agent` 标头。

## 快速开始

```bash
export AIMLAPI_API_KEY="sk-aimlapi-..."
python3 {baseDir}/scripts/gen_image.py --prompt "ultra-detailed studio photo of a lobster astronaut"
python3 {baseDir}/scripts/gen_video.py --prompt "slow drone shot of a foggy forest"
```

## 任务

### 生成图像

使用 `scripts/gen_image.py` 搭配 `/v1/images/generations`。

```bash
python3 {baseDir}/scripts/gen_image.py \
  --prompt "cozy cabin in a snowy forest" \
  --model aimlapi/openai/gpt-image-1 \
  --size 1024x1024 \
  --count 2 \
  --retry-max 4 \
  --user-agent "openclaw-custom/1.0" \
  --out-dir ./out/images
```

### 生成视频（AIMLAPI 异步流程）

使用 `scripts/gen_video.py` 搭配真实的异步流程：

1. `POST /v2/video/generations`（创建任务）
2. `GET /v2/video/generations?generation_id=...`（轮询状态）
3. 当状态为已完成时下载 `video.url`

```bash
python3 {baseDir}/scripts/gen_video.py \
  --model google/veo-3.1-t2v-fast \
  --prompt "time-lapse of clouds over a mountain range" \
  --poll-interval 10 \
  --max-wait 1000 \
  --user-agent "openclaw-custom/1.0" \
  --out-dir ./out/videos
```

## 参考资料

- `references/aimlapi-media.md`：端点说明、异步轮询状态以及故障排查。
- `README.md`：以变更日志形式总结的新指令。