Vertex AI Gateway

Google Vertex AI (Gemini) → OpenAI 兼容接口，运行在 Cloudflare Workers 上。

API Key 已配置服务账号已配置

可用端点

POST https://vertex-gateway.3171711374.workers.dev/v1/chat/completions — 对话（支持 stream: true）
POST https://vertex-gateway.3171711374.workers.dev/v1/completions — 文本补全（旧版接口）
POST https://vertex-gateway.3171711374.workers.dev/v1/embeddings — 文本向量
GET https://vertex-gateway.3171711374.workers.dev/v1/models — 模型列表
GET https://vertex-gateway.3171711374.workers.dev/health — 健康检查（无需鉴权）
POST https://vertex-gateway.3171711374.workers.dev/v1/vertex/generate — Vertex 原生透传

curl 示例

curl https://vertex-gateway.3171711374.workers.dev/v1/chat/completions \
  -H "Authorization: Bearer $PROXY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-2.5-flash",
    "messages": [{"role":"user","content":"你好，用一句话介绍你自己"}]
  }'

OpenAI SDK 示例

from openai import OpenAI

client = OpenAI(
    base_url="https://vertex-gateway.3171711374.workers.dev/v1",
    api_key="你的 PROXY_API_KEY",
)

resp = client.chat.completions.create(
    model="gemini-2.5-flash",
    messages=[{"role": "user", "content": "你好"}],
)
print(resp.choices[0].message.content)

指定区域

默认区域 global。可在模型名后加 @区域，或使用请求头 x-vertex-location 覆盖，例如：

"model": "gemini-2.5-pro@europe-west1"