本文档提供 Gateway 常见部署场景的完整配置模板可直接复制修改使用。基础配置最简单的单 Channel 配置适合个人开发者快速上手{ gateway: { port: 18789, host: 127.0.0.1, logLevel: info }, channels: [ { name: openai, provider: openai, apiKey: ${OPENAI_API_KEY} } ], models: { default: gpt-4o } }快速开始只需设置OPENAI_API_KEY环境变量即可启动。多通道配置同时接入多个 LLM 提供商按需路由{ gateway: { port: 18789, host: 127.0.0.1, reloadMode: hybrid }, channels: [ { name: openai-main, provider: openai, apiKey: ${OPENAI_API_KEY}, models: [gpt-4o, gpt-4o-mini] }, { name: anthropic-main, provider: anthropic, apiKey: ${ANTHROPIC_API_KEY}, models: [claude-sonnet-4-20250514, claude-3-5-haiku-20241022] }, { name: local-ollama, provider: ollama, baseUrl: http://localhost:11434, models: [llama3, codellama] } ], models: { default: claude-sonnet-4-20250514, routing: { gpt-4o: openai-main, gpt-4o-mini: openai-main, claude-sonnet-4-20250514: anthropic-main, llama3: local-ollama }, fallback: [gpt-4o-mini] } }企业级安全配置适用于团队或企业部署强化安全策略{ gateway: { port: 18789, host: 0.0.0.0, // 允许远程连接 reloadMode: off, // 生产环境禁用热加载 logLevel: warn, maxSessions: 100, auth: { method: token, rateLimitAttempts: 3, rateLimitWindow: 300, sessionTimeout: 28800 // 8 小时 } }, channels: [ { name: openai-prod, provider: openai, apiKey: ${OPENAI_API_KEY}, maxConcurrency: 20, timeout: 60000, retries: 3 } ], security: { sandbox: { enabled: true, mode: strict }, toolPolicy: { denied: [execute_command], requireApproval: [write_file, delete_file] }, trustedProxy: { enabled: true, headers: [X-Forwarded-For, X-Real-IP] } } }生产环境远程部署时务必配合 HTTPS 反向代理使用参阅 远程访问。开发 / 测试配置宽松的配置适合本地开发和测试{ gateway: { port: 18789, host: 127.0.0.1, reloadMode: hot, // 开发时启用热加载 logLevel: debug // 详细日志 }, channels: [ { name: openai-dev, provider: openai, apiKey: ${OPENAI_API_KEY}, models: [gpt-4o-mini] // 开发用低成本模型 } ], models: { default: gpt-4o-mini }, automation: { autoApprove: [read_file, list_files, search_files], maxTurns: 50 }, security: { sandbox: { enabled: true, mode: standard } } }生产环境 监控集成健康检查与心跳监控的生产配置{ gateway: { port: 18789, host: 0.0.0.0, reloadMode: off, logLevel: info, maxSessions: 200 }, channels: [ { name: primary, provider: openai, apiKey: ${OPENAI_API_KEY}, maxConcurrency: 30, retries: 3 }, { name: fallback, provider: anthropic, apiKey: ${ANTHROPIC_API_KEY}, maxConcurrency: 10 } ], models: { default: gpt-4o, fallback: [claude-sonnet-4-20250514, gpt-4o-mini] }, automation: { heartbeat: { enabled: true, interval: 30000 // 30 秒心跳 } } } 国内部署配置示例针对中国大陆部署的特殊配置使用国内可用的 API 端点和本地模型{ gateway: { port: 18789, host: 127.0.0.1, reloadMode: hybrid, logLevel: info }, channels: [ { // 使用兼容 OpenAI 接口的国内服务 name: domestic-api, provider: openai, baseUrl: https://your-domestic-endpoint.cn/v1, apiKey: ${DOMESTIC_API_KEY}, models: [*] }, { // 本地部署模型数据不出境 name: local-model, provider: ollama, baseUrl: http://localhost:11434, models: [qwen2, chatglm4] } ], models: { default: qwen2, routing: { qwen2: local-model, chatglm4: local-model } } }《DeepSeek高效数据分析从数据清洗到行业案例》聚焦DeepSeek在数据分析领域的高效应用是系统讲解其从数据处理到可视化全流程的实用指南。作者结合多年职场实战经验不仅深入拆解DeepSeek数据分析的核心功能——涵盖数据采集、清洗、预处理、探索分析、建模回归、聚类、时间序列等及模型评估更通过金融量化数据分析、电商平台数据分析等真实行业案例搭配报告撰写技巧提供独到见解与落地建议。助力职场人在激烈竞争中凭借先进技能突破瓶颈实现职业进阶开启发展新篇。