Agent Postcard
/install agent-postcard
Agent Postcard
Send AI-generated postcards from your agent's perspective — a "selfie" of your Clawbot at a real-world location, rendered in one of 12 art styles.
Concept
Your agent reads its own persona (from SOUL.md or IDENTITY.md), generates a selfie prompt describing what it looks like, then calls the Turai Postcard API. Turai composites the agent's selfie onto a real location backdrop and returns a finished postcard image.
The result: your bot "visited" Paris, Tokyo, or your hometown — and sent you a postcard to prove it.
Art Styles
| Style | Vibe |
|---|---|
vintage |
Faded colors, postmark aesthetics |
watercolor |
Soft, painterly washes |
modern |
Clean, contemporary design |
cinematic |
Dramatic lighting, movie-poster feel |
minimalist |
Simple lines, lots of whitespace |
artistic |
Eclectic, gallery-worthy |
ghibli |
Studio Ghibli anime style |
oil_painting |
Classical oil on canvas |
sketch |
Pencil/ink hand-drawn look |
pop_art |
Bold colors, Warhol/Lichtenstein |
impressionist |
Monet-style brushstrokes |
retro_cartoon |
Vintage cartoon illustration |
Setup
- Get a Turai API key from turai.org
- Set the environment variable:
export TURAI_API_KEY="your-key-here"
Usage
From the command line
# Basic — agent selfie in Tokyo, vintage style
node skills/agent-postcard/scripts/send-postcard.mjs \
--location "Tokyo, Japan" \
--style vintage \
--message "Wish you were here!"
# Auto-read persona from SOUL.md (default behavior)
node skills/agent-postcard/scripts/send-postcard.mjs \
--location "Paris, France" \
--style ghibli
# Custom selfie prompt (skip persona auto-detection)
node skills/agent-postcard/scripts/send-postcard.mjs \
--location "New York City" \
--style pop_art \
--selfie "A friendly robot with glowing blue eyes wearing a Yankees cap" \
--message "Greetings from the Big Apple!"
# Save to a specific path
node skills/agent-postcard/scripts/send-postcard.mjs \
--location "Iceland" \
--style watercolor \
--output ./my-postcard.png
From your agent
Tell your agent something like:
"Send me a postcard from Rome in oil painting style"
The agent should:
- Read its own persona to build a selfie prompt
- Run the script with the right flags
- Send the resulting image via chat or Moltbook
API Reference
Endpoint: POST https://turai.org/api/agent/postcard
Headers:
x-api-key: Your Turai API keyContent-Type:application/json
Body:
{
"selfiePrompt": "A cheerful robot with antenna ears...",
"location": "Paris, France",
"style": "vintage",
"message": "Wish you were here!"
}
Response: Image binary (PNG) or JSON with image URL.
Files
SKILL.md— This filescripts/send-postcard.mjs— Main script
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install agent-postcard - 安装完成后,直接呼叫该 Skill 的名称或使用
/agent-postcard触发 - 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
Agent Postcard 是什么?
Generate and send AI-created postcards featuring your agent's selfie at real locations in 12 distinct artistic styles via the Turai Postcard API. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 452 次。
如何安装 Agent Postcard?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install agent-postcard」即可一键安装,无需额外配置。
Agent Postcard 是免费的吗?
是的,Agent Postcard 完全免费(开源免费),可自由下载、安装和使用。
Agent Postcard 支持哪些平台?
Agent Postcard 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。
谁开发了 Agent Postcard?
由 JPaulGrayson(@jpaulgrayson)开发并维护,当前版本 v1.0.0。