Vellum 是一个面向生产级 AI 应用的 LLM 开发平台，提供提示词管理、模型网关、评测（evals）和可视化工作流编排。团队用它统一管理多个大模型供应商（OpenAI、Anthropic、Gemini 等），一键切换模型并对比效果，还能做回归测试与 A/B 测试，是 MLOps 团队构建 RAG 与 Agent 应用的常用选择。 🔗 文章: /zh/post/self-hosted-ai-agent-framework-comparison
Vellum is an LLM development platform for production AI apps — prompt management, a unified model gateway, evals and visual workflow orchestration across OpenAI, Anthropic, Gemini and more. A go-to for MLOps teams building RAG and agent applications. 🔗 Article: /post/self-hosted-ai-agent-framework-comparison