export default {
async fetch(request, env, ctx) {
// 处理 CORS 预检请求
if (request.method === 'OPTIONS') {
return new Response(null, {
status: 200,
headers: {
'Access-Control-Allow-Origin': '*',
'Access-Control-Allow-Methods': 'GET, POST, OPTIONS',
'Access-Control-Allow-Headers': 'Content-Type, Authorization',
'Access-Control-Max-Age': '86400',
},
});
}
// 验证 API Key
const authHeader = request.headers.get('Authorization');
if (!authHeader || !authHeader.startsWith('Bearer ')) {
return new Response(JSON.stringify({
error: { message: 'API key required', type: 'invalid_request_error', code: 'invalid_api_key' }
}), { status: 401, headers: { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' } });
}
const apiKey = authHeader.substring(7);
const validApiKeys = env.VALID_API_KEYS ? env.VALID_API_KEYS.split(',') : ['your-api-key-here'];
if (!validApiKeys.includes(apiKey)) {
return new Response(JSON.stringify({
error: { message: 'Invalid API key', type: 'invalid_request_error', code: 'invalid_api_key' }
}), { status: 401, headers: { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' } });
}
const url = new URL(request.url);
// 模型映射
const modelMap = {
'deepseek-r1': '@cf/deepseek-ai/deepseek-r1-distill-qwen-32b',
'gpt-oss-120b': '@cf/openai/gpt-oss-120b',
'gpt-oss-20b': '@cf/openai/gpt-oss-20b',
'llama-4-scout': '@cf/meta/llama-4-scout-17b-16e-instruct',
'qwen2.5-coder': '@cf/qwen/qwen2.5-coder-32b-instruct',
'gemma-3': '@cf/google/gemma-3-12b-it'
};
// 聊天接口
if (url.pathname === '/v1/chat/completions' && request.method === 'POST') {
try {
const body = await request.json();
if (!body.messages || !Array.isArray(body.messages)) {
return new Response(JSON.stringify({
error: { message: 'Messages must be an array', type: 'invalid_request_error', code: 'invalid_parameter' }
}), { status: 400, headers: { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' } });
}
const model = body.model || 'deepseek-r1';
const cfModel = modelMap[model];
if (!cfModel) {
return new Response(JSON.stringify({
error: { message: `Model '${model}' not supported`, type: 'invalid_request_error', code: 'model_not_found' }
}), { status: 400, headers: { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' } });
}
// 构造 AI 请求参数
let aiRequest = {};
let useResponsesAPI = cfModel.startsWith('@cf/openai/gpt-oss');
if (useResponsesAPI) {
// Responses API 格式
const systemMsg = body.messages.find(m => m.role === 'system')?.content || "You are a helpful assistant.";
const userMsgs = body.messages.filter(m => m.role === 'user').map(m => m.content).join("\n");
aiRequest = {
input: userMsgs,
instructions: systemMsg,
temperature: body.temperature ?? 0.7,
top_p: body.top_p ?? 0.9,
max_tokens: body.max_tokens ?? 2048,
reasoning: body.reasoning ?? { effort: "medium" }
};
} else {
// 旧模型：拼接 prompt
let prompt = '';
for (const message of body.messages) {
if (message.role === 'system') prompt += `System: ${message.content}\n\n`;
if (message.role === 'user') prompt += `User: ${message.content}\n\n`;
if (message.role === 'assistant') prompt += `Assistant: ${message.content}\n\n`;
}
prompt += 'Assistant: ';
aiRequest = {
prompt,
temperature: body.temperature ?? 0.7,
top_p: body.top_p ?? 0.9,
max_tokens: body.max_tokens ?? 4096,
};
}
// 调用 Cloudflare AI
const response = await env.AI.run(cfModel, aiRequest);
const completionId = 'chatcmpl-' + Math.random().toString(36).substring(2, 15);
const timestamp = Math.floor(Date.now() / 1000);
// 获取最终回答内容
let assistantContent = "";
if (useResponsesAPI) {
if (response.output && Array.isArray(response.output)) {
assistantContent = response.output
.flatMap(msg => msg.content
.filter(c => c.type === "output_text")
.map(c => c.text)
)
.join("\n");
}
} else {
assistantContent = response.response ?? "";
}
// 流式输出
if (body.stream) {
const encoder = new TextEncoder();
const stream = new ReadableStream({
start(controller) {
// 开始事件
controller.enqueue(encoder.encode(`data: ${JSON.stringify({
id: completionId,
object: 'chat.completion.chunk',
created: timestamp,
model,
choices: [{ index: 0, delta: { role: 'assistant', content: "" }, finish_reason: null }]
})}\n\n`));
// 模拟逐块输出
const chunkSize = 20;
for (let i = 0; i < assistantContent.length; i += chunkSize) {
const chunk = assistantContent.slice(i, i + chunkSize);
controller.enqueue(encoder.encode(`data: ${JSON.stringify({
id: completionId,
object: 'chat.completion.chunk',
created: timestamp,
model,
choices: [{ index: 0, delta: { content: chunk }, finish_reason: null }]
})}\n\n`));
}
// 结束事件
controller.enqueue(encoder.encode(`data: ${JSON.stringify({
id: completionId,
object: 'chat.completion.chunk',
created: timestamp,
model,
choices: [{ index: 0, delta: {}, finish_reason: 'stop' }]
})}\n\n`));
controller.enqueue(encoder.encode('data: [DONE]\n\n'));
controller.close();
}
});
return new Response(stream, {
headers: { 'Content-Type': 'text/event-stream', 'Cache-Control': 'no-cache', 'Connection': 'keep-alive', 'Access-Control-Allow-Origin': '*' },
});
}
// 非流式输出
const chatCompletion = {
id: completionId,
object: 'chat.completion',
created: timestamp,
model,
choices: [{ index: 0, message: { role: 'assistant', content: assistantContent }, finish_reason: 'stop' }],
usage: {
prompt_tokens: Math.ceil(JSON.stringify(body.messages).length / 4),
completion_tokens: Math.ceil(assistantContent.length / 4),
total_tokens: Math.ceil((JSON.stringify(body.messages).length + assistantContent.length) / 4)
}
};
return new Response(JSON.stringify(chatCompletion), { headers: { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' } });
} catch (error) {
console.error('Error:', error);
return new Response(JSON.stringify({
error: { message: 'Internal server error', type: 'server_error', code: 'internal_error' }
}), { status: 500, headers: { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' } });
}
}
// 模型列表
if (url.pathname === '/v1/models' && request.method === 'GET') {
const models = Object.keys(modelMap).map(id => ({
id,
object: 'model',
created: Math.floor(Date.now() / 1000),
owned_by: 'cloudflare',
permission: [{ id: 'modelperm-' + id, object: 'model_permission', created: Math.floor(Date.now() / 1000), allow_create_engine: false, allow_sampling: true, allow_logprobs: false, allow_search_indices: false, allow_view: true, allow_fine_tuning: false, organization: '*', group: null, is_blocking: false }]
}));
return new Response(JSON.stringify({ object: 'list', data: models }), { headers: { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' } });
}
// 健康检查
if (url.pathname === '/health' && request.method === 'GET') {
return new Response(JSON.stringify({ status: 'healthy', timestamp: new Date().toISOString(), models: Object.keys(modelMap) }), { headers: { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' } });
}
// 404
return new Response(JSON.stringify({
error: { message: 'Not found', type: 'invalid_request_error', code: 'not_found' }
}), { status: 404, headers: { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' } });
},
};
需要绑定workers ai，变量名为AI
添加api key变量VALID_API_KEYS，多个key 逗号分割
功能特性
-
OpenAI API 兼容格式 -
API 密钥认证 -
流式响应 (gpt-oss的模型不支持) -
多模型支持 -
Cloudflare workers 部署
支持的模型
| 模型名称 | Cloudflare 对应模型 |
|---|---|
deepseek-r1 |
@cf/deepseek-ai/deepseek-r1-distill-qwen-32b |
gpt-oss-120b |
@cf/openai/gpt-oss-120b |
gpt-oss-20b |
@cf/openai/gpt-oss-20b |
llama-4-scout |
@cf/meta/llama-4-scout-17b-16e-instruct |
qwen2.5-coder |
@cf/qwen/qwen2.5-coder-32b-instruct |
gemma-3 |
@cf/google/gemma-3-12b-it |
响应格式
遵循 OpenAI Chat Completions API 格式，包含完整的 tokens 使用统计。
示例请求
curl -X POST "https://your-worker.workers.dev/v1/chat/completions" \
-H "Authorization: Bearer your-api-key" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek-r1",
"messages": [
{"role": "user", "content": "你好，请介绍一下自己"}
],
"stream": false
}'
响应
{
"id": "chatcmpl-lt3fpzwy6vg",
"object": "chat.completion",
"created": 1758553310,
"model": "deepseek-r1",
"choices": [
{
"index": 0,
"message": {
"role": "assistant",
"content": "你好！我是DeepSeek-R1，一个由深度求索公司开发的智能助手，我会尽我所能为您提供帮助。\n</think>\n\n你好！我是DeepSeek-R1，一个由深度求索公司开发的智能助手，我会尽我所能为您提供帮助。"
},
"finish_reason": "stop"
}
],
"usage": {
"prompt_tokens": 10,
"completion_tokens": 27,
"total_tokens": 37
}
}